IJACSA Volume 13 Issue 7 - thesai.org

IJACSA Volume 13 Issue 7

Copyright Statement: This is an open access publication licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, even commercially as long as the original work is properly cited.

View Full Issue

Paper 1: Research Progress and Trend of the Machine Learning based on Fusion

Abstract: Machine learning is widely used in the data processing including data classification, data regression, data mining and so on, and based on a single type of machine learning technology, it is often difficult to meet the requirements of data processing; in recent years, the machine learning based on fusion has become an important approach to improve data processing effect, and at the same time, corresponding summary study is relatively limited. In this study, we summarize and compare different types of fusion machine learning such as ensemble learning, federated learning and transfer learning from the perspectives of classification, principle and characteristics, and try to explore the research development trend, in order to provide effective reference for subsequent related research and application; furthermore, as an application of fusion machine learning，we also conduct a study on the modeling optimization for car service complaint text classification.

Author 1: Chen Xiao Yu

Author 2: Zhang Xiao Min

Author 3: Song Ying

Author 4: Gao Feng

Keywords: Machine learning; fusion; ensemble learning; federated learning; transfer learning

Paper 2: Detection of Premature Ventricular Contractions using 12-lead Dynamic ECG based on Squeeze-Excitation Residual Network

Abstract: Premature ventricular contraction (PVC) is a very common arrhythmia that can originate in any part of the ventricle and is one of the important causes of sudden cardiac death. Timely and rapid detection of PVC on dynamic electrocardiogram (ECG) recording for patients with cardiovascular diseases is of great significance for clinical diagnosis. Furthermore, it can facilitate the planning and execution of radiofrequency ablation. But the dynamic ECGs can be easily contaminated by various noises and its morphological characteristics show significant variations for different patients. Though the deep learning methods achieved outstanding performance in ECG automatic recognition, there are still some limitations, such as overfitting, gradient disappearance or gradient explosion in deep networks. Therefore, a residual module is constructed using the squeeze-excitation method to alleviate the problems. A 20-layer squeeze-extraction residual network (SE-ResNet) containing multiple squeeze-extraction modules was designed for real-time PVC detection on 12-lead dynamic ECG. The algorithm was evaluated using the dynamic 12-lead ECGs in INCART database (168,379 heartbeats in total). The experimental results show that the test accuracy of the method proposed in this paper is 98.71%, and the specificity and sensitivity of PVC are 99.12% and 99.59%, respectively. Under the same dataset and experimental platform, the average recognition accuracy of our proposed method is increased by 0.73%, 1.55%, 2.9% and 1.65% compared with the results obtained by CNN, Inception, AlexNet and deep multilayer perceptron, respectively. The proposed scheme provides a new method for real-time detection of PVC on dynamic 12-lead ECGs. The experiment results show that the proposed method outperforms state-of-the-art methods, and has good potential for clinical applications.

Author 1: Duan Li

Author 2: Tingting Sun

Author 3: Yibai Xue

Author 4: Yilin Xie

Author 5: Xiaolei Chen

Author 6: Jiaofen Nan

Keywords: Dynamic ECG; squeeze-excitation; residual network; premature ventricular contraction

Paper 3: Personality Classification Model of Social Network Profiles based on their Activities and Contents

Abstract: Social networks have become an important part of everyday life, especially after the latest technologies such as smartphones, tablets, and laptops have become widespread. Individuals spend a lot of time on social media and express their feelings and opinions through statuses, comments, and updates which could be a way to understand and classify their personalities. The personalities in psychological science are divided into five classes according to the Big-five model (Openness, Extraversion, Consciousness, Agreeableness, and Neurotic). This model shows the key features with their weights for each personality. In this paper, a proposed model is developed for detecting the personality features from users’ activities in social networks. In this model, machine learning techniques are used for predicting the personalities with a score for each Big-five model type and sorting them in descending order. The personality classification model will be useful in developing a better understanding of the user profile and specifically targeting users with appropriate advertising. Any social media network user's personality can be predicted by using their posts and status updates to get better accuracy. The experimental results of the model in this study provide an enhancement because it can predict the precise score of one user in each factor of the Big-five. The proposed model was tested on a dataset extracted from Facebook and manually classified by experts, and it achieved 89.37% accuracy.

Author 1: Mervat Ragab Bakry

Author 2: Mona Mohamed Nasr

Author 3: Fahad Kamal Alsheref

Keywords: Psychological personality; machine learning techniques; big-five; LinearSVC

Paper 4: Customer Profiling Method with Big Data based on BDT and Clustering for Sales Prediction

Abstract: We propose a method for customer profiling based on Binary Decision Tree: BDT and k-means clustering with customer related big data for sales prediction; valuable customer findings as well as customer relation improvements. Through the customer related big data, not only sales prediction but also categorization of customers as well as Corporate Social Responsibility (CSR) can be done. This paper describes a method for these purposes. Examples of the analyzed data relating to the sales prediction, valuable customer findings and customer relation improvements are shown here. It is found that the proposed method allows sales prediction, valuable customer findings with some acceptable errors.

Author 1: Kohei Arai

Author 2: Zhan Ming Ming

Author 3: Ikuya Fujikawa

Author 4: Yusuke Nakagawa

Author 5: Tatsuya Momozaki

Author 6: Sayuri Ogawa

Keywords: Customer profiling; binary decision tree: BDT; corporate social responsibility (CSR); k-means clustering; sales prediction; valuable customer findings

Paper 5: A New Hybrid-Heuristic Approach for Vertex p-Median Location Problems

Abstract: In this paper, a new hybridization of a Myopic and Neighborhood approaches is proposed to solve large-size vertex p-median location problems. The effectiveness and efficiency of our approach are demonstrated empirically through an intensive computational experiment on large-size instances taken from TSPLib and BIRCH datasets, with the number of nodes varying from 734 to 9,976 for the former and from 9,600 to 20,000 nodes for the latter. The results show that the new approach, though relatively simple, yields better solutions compared to the ones in the literature. This demonstrates that a simpler approach that takes into account the advantages of other methods can lead to promising outcome and has the potential of being adopted in other combinatorial optimization problems.

Author 1: Hassan Mohamed Rabie

Author 2: Said Salhi

Keywords: P-median; discrete location problems; myopic heuristic; neighborhood heuristic

Paper 6: Annotated Corpus with Negation and Speculation in Arabic Review Domain: NSAR

Abstract: Negation and speculation detection are critical for Natural Language Processing (NLP) tasks, such as sentiment analysis, information retrieval, and machine translation. This paper presents the first Arabic corpus in the review domain annotated with negation and speculation. The Negation and Speculation Arabic Review (NSAR) corpus consists of 3K randomly selected review sentences from three well-known and benchmarked Arabic corpora. It contains reviews from different categories, including books, hotels, restaurants, and other products written in various Arabic dialects. The negation and speculation keywords have been annotated along with their linguistic scope based on the annotation guidelines reviewed by an expert linguist. The inter-annotator agreement between two independent annotators, Arabic native speakers, is measured using the Cohen’s Kappa coefficients with values of 95 and 80 for negation and speculation, respectively. Furthermore, 29% of this corpus includes at least one negation instance, while only 4% of this corpus contains speculative content. Therefore, the Arabic reviews focus more on negation structures rather than speculation. This corpus will be available for the Arabic research community to handle these critical phenomena.

Author 1: Ahmed Mahany

Author 2: Heba Khaled

Author 3: Nouh Sabri Elmitwally

Author 4: Naif Aljohani

Author 5: Said Ghoniemy

Keywords: Arabic NLP; negation; speculation; uncertainty; annotation; annotation guidelines; corpus; review domain; sentiment analysis

Paper 7: Drought Prediction and Validation for Desert Region using Machine Learning Methods

Abstract: Drought prediction serves as an early warning to the effective management of water resources to avoid the drought impact. The drought prediction is carried out for arid, semi-arid, sub-humid, and humid climate types in the desert region. The drought is predicted using Standardized precipitation evapotranspiration index (SPEI). The application of machine learning methods such as artificial neural network (ANN), K-Nearest Neighbors (KNN), and Deep Neural Network (DNN) for the drought prediction suitability is analyzed. The SPEI is predicted using the aforesaid machine learning methods with inputs used to calculate SPEI. The predictions are assessed using statistical indicators. The coefficient of determination of ANN, KNN, and DNN are 0.93, 0.83, and 0.91 respectively. The mean square error of ANN, KNN, and DNN are 0.065, 0.512, and 0.52 respectively. The mean absolute error of ANN, KNN, and DNN are 0.001, 0.512, and 0.01 respectively. Based on results of statistical indicator and validations it is found that DNN is suitable to predict drought in all the four types of desert region.

Author 1: Azmat Raja

Author 2: Gopikrishnan T

Keywords: Drought; SPEI; machine learning; water resources; prediction

Paper 8: An SDN-based Decision Tree Detection (DTD) Model for Detecting DDoS Attacks in Cloud Environment

Abstract: Detecting Distributed Denial of Service (DDoS) attacks has become a significant security issue for various network technologies. This attack has to be detected to increase the system’s reliability. Though various traditional studies are present, they suffer from data shift issues and accuracy. Hence, this study intends to detect DDoS attacks by classifying the normal and malicious traffic. The study aims to solve the data shift issues by using the introduced Decision Tree Detection (DTD) model encompassing of Greedy Feature Selection (GFS) algorithm and Decision Tree Algorithm (DTA). It also attempts to enhance the proposed model’s detection rate (accuracy) above 90%. Various processes are involved in DDoS attack detection. Initially, the gureKddcup dataset is loaded to perform pre-processing. This process is essential for removing noisy data. After this, feature selection is performed to select only the relevant features, removing the irrelevant data. This is then fed into the train and test split. Following this, Software Defined Networking (SDN) based DTA is used to classify the normal and malicious traffic, then given to the trained model for predicting this attack. Performance analysis is undertaken by comparing the proposed model with existing systems in terms of accuracy, MCC (Matthew’s Correlation Coefficient), sensitivity, specificity, error rate, FAR (False Alarm Rate), and AUC (Area under Curve). This analysis is carried out to evaluate the efficacy of the proposed model, which is verified through the results.

Author 1: Jeba Praba. J

Author 2: R. Sridaran

Keywords: Distributed denial of service attack; greedy feature selection; decision tree algorithm; software defined networking; cloud and decision tree detection

Paper 9: Mono Camera-based Human Skeletal Tracking for Squat Exercise Abnormality Detection using Double Exponential Smoothing

Abstract: Human action analysis is an enthralling area of research in artificial intelligence, as it may be used to improve a range of applications, including sports coaching, rehabilitation, and monitoring. By forecasting the body's vital position of posture, human action analysis may be performed. Human body tracking and action recognition are the two primary components of video-based human action analysis. We present an efficient human tracking model for squat exercises using the open-source MediaPipe technology. The human posture detection model is used to detect and track the vital body joints within the human topology. A series of critical body joint motions are being observed and analysed for aberrant body movement patterns while conducting squat workouts. The model is validated using a squat dataset collected from ten healthy people of varying genders and physiques. The incoming data from the model is filtered using the double exponential smoothing method;the Mean Squared Error between the measured and smoothed angles is determined to classify the movement as normal or abnormal. Level smoothing and trend control have parameters of 0.8928 and 0.77256, respectively. Six out of ten subjects in the trial were precisely predicted by the model. The mean square error of the signals obtained under normal and abnormal squat settings is 56.3197 and 29.7857, respectively. Thus, by utilising a simple threshold method, the low-cost camera-based squat movement condition detection model was able to detect the abnormality of the workout movement.

Author 1: Muhammad Nafis Hisham

Author 2: Mohd Fadzil Abu Hassan

Author 3: Norazlin Ibrahim

Author 4: Zalhan Mohd Zin

Keywords: Abnormality movement; double exponential smoothing; skeletal tracking; mediapipe; squat exercise

Paper 10: Solving the Imbalanced and Limited Data Labeled for Automated Essay Scoring using Cost Sensitive XGBoost and Pseudo-Labeling

Abstract: There are two main problems on forming the Automatic Essay Scoring Model. They are the datasets having imbalanced amount of the right and wrong answers and the minimal use of labeled data in the model training. The model forming based on these problems is divided into three main points, namely word representation, Cost-Sensitive XGBoost Classification, and adding unlabeled data with the Pseudo-Labeling Technique. The essay answer data is converted into a vector using the trained word vector fastText. Furthermore, the classification of unlabeled data was carried out using the Cost-Sensitive XGBoost Method. The data labeled by the classification model is added as training data for the new classification model form. The process is carried out iteratively. This research is about using the combination of Cost-Sensitive XGBoost Classification and Pseudo-Labeling which is expected to solve the problems. For the 0th iteration, the dataset having a ratio of the amount of "right" labeled data with the amount of "right" labeled data is close to 1, in other words a balanced dataset or a ratio that is more than 1 produces a model with better performance. Thus, the selection of training data at an early stage must pay attention to this ratio. In addition, the use of the Hybrid Method on these datasets can save labeled data 56 times compared to the AdaBoost Method. Hybrid model is able to produce F1-Measure more than 95.6%, so it can be concluded that the Hybrid Method, which combines the XGBoost and Pseudo-Labeling Cost-Sensitive Classification with Self Training, is able to overcome the problem of unbalanced datasets and data limited label.

Author 1: Marvina Pramularsih

Author 2: Mardhani Riasetiawan

Keywords: Imbalanced data; limited labeled data; automated essay scoring; cost sensitive XGBoost; pseudo-labeling

Paper 11: Prediction of Instructor Performance using Machine and Deep Learning Techniques

Abstract: The quality of instructors’ performance mainly influences the quality of educational services in higher educational institutions. One of the major challenges of higher educational institutions is the accumulated amount of data and how it can be utilized to boost the academic programs quality. The recent advancements in Artificial Intelligence techniques, including machine and deep learning models, have led to the expansion in practical prediction for various fields. In this paper, a dataset was collected from UCI Repository, University of California, for the prediction of instructor performance. In order to find how effective the instructor in the higher education systems is, a group of machine and deep learning algorithms were applied to predict instructor performance in higher education systems. The best machine-learning algorithm was Extra Trees Regressor with Accuracy (98.78%), Precision (98.78%), Recall (98.78%), F1-score (98.78%); however, the proposed deep learning algorithm achieved Accuracy (98.89%), Precision (98.91%), Recall (98.94%), and F1-score (98.92%).

Author 1: Basem S. Abunasser

Author 2: Mohammed Rasheed J. AL-Hiealy

Author 3: Alaa M. Barhoom

Author 4: Abdelbaset R. Almasri

Author 5: Samy S. Abu-Naser

Keywords: Education; deep learning; machine learning; prediction; instructor performance

Paper 12: Application-based Usability Evaluation Metrics

Abstract: Testing is one of the vital stages in the software development life cycle (SDLC). Usability testing is a very important field that helps the applications be easily used by the end-users. Because of the importance of usability testing, a metrics has been developed to help in measuring the usability through converting the main qualitative usability attributes in ISO to quantitative steps that provide the developer a framework to follow in developing to achieve usability of their applications and helps the tester with a checklist and a tool to measure the usability percentage of their application. The framework provides a set of steps to achieve the usability attributes and answers the question of how you could measure this attribute with the defined steps. The framework results in a 95% average accuracy in the high-rate application and a 59% average accuracy in the low-rate application. Finally, the framework is programmed in a tool to measure the usability percentage of the application through a checklist and provides a scheme to help the developer achieve the best results in usability.

Author 1: Hanaa Bayomi

Author 2: Noura A.Sayed

Author 3: Hesham Hassan

Author 4: Khaled Wassif

Keywords: Usability; human-computer interaction; evaluation; quantitative attributes; testing

Paper 13: Design and Development of Face Mask Reminder Box Technology using Arduino Uno

Abstract: The World Health Organization (WHO) declared the COVID-19 pandemic on 12 Mar, 2020, due to the growth in the number of cases worldwide. WHO advises wearing a face mask and practicing social distancing, which has played a crucial role in prevention and control measures that can prevent the spread of COVID-19. Thus, this paper presents the process through which face mask box is equipped with a voice reminder and sensor. It is made with the help of an Arduino Uno board to give awareness or reminder whenever a person is alerted with a voice reminder to wear a face mask before going outside. It can be helpful, especially in the pandemic era, as a new norm of practice in wearing a mask.

Author 1: Chee Ken Nee

Author 2: Rafiza Abdul Razak

Author 3: Muhamad Hariz Bin Muhamad Adnan

Author 4: Wan Fatin Liyana Mutalib

Author 5: Nur Fatirah Roslee

Author 6: Noor Mursheeda Mahyuddin

Keywords: Face mask box; COVID-19; Voice reminder; Arduino; New norm

Paper 14: An Efficient Unusual Event Tracking in Video Sequence using Block Shift Feature Algorithm

Abstract: The area of video technology is rapidly growing owing to advancements in intelligent video systems in sensor operations, higher bandwidth capacity, storage, and high-resolution displays. This led to the proliferation of video-based computing modeling to perform specific tasks on video sequences to gain more insight from the data. Visual tracking of events is a core component in video visual surveillance systems that classify and track moving objects to describe their behavioral aspects. The prime motive behind intelligent video systems is to perform efficient video analytics to meet the specific requirements of the user/use-cases. It involves a self-directed paradigm to understand event sequences, reducing the computational burden of characterizing the activities. The study incorporates a block-shift feature algorithm and introduces a novel computational research method for unusual event tracking in video sequences. The formulated approach employs a framework combining operational blocks to compute sequential operations such as block-matching from the dictionary of motion estimations. Before applying the learning model, the subsequent analysis procedure adds feature lexicon and dominant attributes to make the execution computationally efficient. Further, it uses a sparse-non negative factorization approach to organize the informative details into k possible finite clusters. The event detection outcome from the training datasets of video sequences shows better experimental results than the traditional highly cited related approach of unusual object detection and tracking.

Author 1: Karanam Sunil Kumar

Author 2: N P Kavya

Keywords: Object detection; tracking; learning models; video sequence analysis

Paper 15: An Ontological Model based on Machine Learning for Predicting Breast Cancer

Abstract: Breast cancer is mostly a female disease, but it may affect men as well even at a considerably lower percentage. An automated diagnosis system should be built for early detection because manual breast cancer diagnosis takes a long time. Doctors have lately achieved significant advances in the early identification and treatment of breast cancer in order to decrease the rate of mortality caused by the latter. Researchers, on the other hand, are analysing large amounts of complicated medical data by employing a combination of statistical and machine learning methodologies to assist clinicians in predicting breast cancer. Various machine learning approaches, including ontology-based Machine Learning methods, have lately played an essential role in medical science by building an automated system that can identify breast cancer. This study examines and evaluates the most popular machine learning algorithms, besides the ontological model based on Machine Learning. Among the classification methods investigated were Naive Bayes, Decision Tree, Logistic Regression, Support Vector Machine, Artificial Neural Network, Random Forest, and k-Nearest Neighbours. The dataset utilized has 683 instances and is available for download from the Kaggle website. The findings are assessed using performance measures generated from the confusion matrix, such as F-Measure, Accuracy, Precision, and Recall. The ontology model surpassed all machine learning techniques, according to the results.

Author 1: Hakim El Massari

Author 2: Noreddine Gherabi

Author 3: Sajida Mhammedi

Author 4: Hamza Ghandi

Author 5: Fatima Qanouni

Author 6: Mohamed Bahaj

Keywords: Machine learning; prediction; ontology; semantic web rule language; decision tree; breast cancer

Paper 16: An Intelligent Transport System in VANET using Proxima Analysis

Abstract: There is no proper structure for Vehicular ad hoc networks (VANETs). VANET generates several mobility vehicles that move in different directions by connecting the vehicles and transferring the data between the source and destination which is very useful information. In this system, a small network is created with vehicles and other devices that behave like nodes in the network. Sometimes for better communication, VANET uses suitable hardware for improving the performance of the network. Reliability is one of the significant tasks that perform the needful operations and methods based on the conditions at a specific time. To disturb the VANETS, the attacker tries to hit the server and that causes damage to the server. This paper mainly focused on detecting the falsification nodes by analyzing the behavior of the models. In this paper, an improved intelligent transportation system (ITS) Proxima analysis is introduced to find the accurate falsification nodes. The proposed approach is the integration of KNN and RF with Proxima analysis. The main aim of the Proxima is to analyze the falsification nodes within the network and improve the mobility of the vehicles by sending source to destination without any miscommunication.

Author 1: Satyanarayana Raju K

Author 2: Selvakumar K

Keywords: Vehicular Ad hoc Network (VANET); intelligent transportation system (ITS); KNN; RF

Paper 17: Virtual Reality, a Method to Achieve Social Acceptance of the Communities Close to Mining Projects: A Scoping Review

Abstract: Background: Virtual reality (VR) technology is an effective, interactive and immersive type of communication since it produces greater interest and attention in the user, thereby allowing greater understanding and comprehension than with more traditional methods. On the other hand, not much information is known about the application of this novel technology in the context of social acceptance as far as the mining sector is concerned; our approach and methodology were based on scoping review (Prisma-SrC, Daudt et al., Arksey, and O’Malley). The research terms were also planned before, with the aim of carrying out three posterior screening levels, among which was the use of EndNote 20 and the PICO framework. Exhaustive research was carried out in nine databases. We obtained n=2 research articles of n=923 initially found, all of which went through three levels of filtering. The chosen articles were evaluated according to Hawker et al. 's methodological rigor, to be included in the review. This scoping review could be the starting point for a series of further investigations that would fill the gap in the literature on this topic, emphasizing experimental articles to confirm the impact of virtual reality technologies on the communities within the sphere of influence of a mining project.

Author 1: Patricia López-Casaperalta

Author 2: Jeanette Fabiola Díaz-Quintanilla

Author 3: José Julián Rodríguez-Delgado

Author 4: Alejandro Marcelo Acosta-Quelopana

Author 5: Aleixandre Brian DuchePérez

Keywords: Mining projects; social acceptance; virtual reality; interaction; mining communities

Paper 18: Semantically Query All (Squerall): A Scalable Framework to Analyze Data from Heterogeneous Sources at Different Levels of Granularity

Abstract: Retracted: After careful and considered review of the content of this paper by a duly constituted expert committee, this paper has been found to be in violation of IJACSA`s Publication Principles. We hereby retract the content of this paper. Reasonable effort should be made to remove all past references to this paper.

Author 1: Iqbal Hasan

Author 2: Majid Zaman

Author 3: Sheikh Amir Fayaz

Author 4: Ifra Altaf

Author 5: Muheet Ahmed Butt

Author 6: S.A.M Rizvi

Keywords: Heterogeneous data; data warehouse; big data; presto; spark; squerall

Paper 19: Evaluation on the Effects of 2D Animation as a Verbal Apraxia Therapy for Children with Verbal Apraxia of Speech

Abstract: This article presents an evaluation of 2D Animation of Learning Numbers and Letters for Children with Verbal Apraxia. The developed application provides some knowledge and encourage children with verbal apraxia to learn and know about numbers and letters. An experimental testing was conducted to evaluate the usability of the developed application, aimed as a therapy for the children who suffer this apraxia across all age. Five important evaluation components such as learnability, usability, accessibility, functionality, and effectiveness were included in this testing to investigate the user engagement and satisfaction of the proposed medical and educational learning system. Online questionnaires were distributed as a method to collect user testing outputs. A total of 33 respondents from multimedia designers, practitioners, psychologists, and parents were involved in this survey. The results of the testing indicate that majority of respondents are satisfied with the outcomes of the 2D animation video. The results presented may facilitate improvements in the teaching syllabus for students with speech and language disorder and produce a great visual animation treatment to the users.

Author 1: Muhammad Taufik Hidayat

Author 2: Sarni Suhaila Rahim

Author 3: Shahril Parumo

Author 4: Nurul Najihah A’bas

Author 5: Muhammad ‘Ammar Muhammad Sani

Author 6: Hilmi Abdul Aziz

Keywords: Childhood apraxia of speech; verbal dyspraxia; speech and language disorder; 2D animation; visual animation treatment; evaluation

Paper 20: An Evaluation Method for Service-Oriented Architecture Maturity Model

Abstract: SOA maturity model was used to clarify and provide a common definition of SOA inside an organization. The model provides an abstract overview of SOA adoption by characterizing evolutionary levels. However, this study found that there is a lacking on how the previous models were evaluated to show that the model is conforming to the specification and can be implemented in the real-world environment. Therefore, this study aims to provide the evaluation method for the SOA maturity model through the verification and validation process. The Integrated Adoption Maturity for Service-Oriented Architecture (IAMSOA) model was chosen and the verification process is being performed through expert review where the study identifies the experts, determines the verification criteria, and collects and analyzes the feedback; while the validation was performed through case study by identifying the organization, determining the validation criteria, brainstorming, and collecting and analyzing the feedback. The verification results show that the evaluated model is comprehensive, understandable, accurate, and well-organized. Moreover, the validation results reveal that it is feasible and practical to be executed in the real environment. Conclusively, this study has successfully evaluated one of the SOA maturity models and shows the verification and validation process in detail which can be re-enacted in different projects and settings.

Author 1: Mohd Hamdi Irwan Hamzah

Author 2: Ezak Fadzrin Ahmad Shaubari

Keywords: Maturity model; model evaluation; model validation; model verification; service-oriented architecture

Paper 21: Fake Face Generator: Generating Fake Human Faces using GAN

Abstract: As machine learning is growing rapidly, creating art and images by machine is one of the most trending topics in current time. It has enormous applications in our current day to day life. Various researchers have researched this topic and they try to implement various ideas and most of them are based on CNN or other tools. The aim of our work is to generate comparatively better real-life fake human faces with low computational power and without any external image classifier, rather than removing all the noise and maximizing the stabilization which was the main challenge of the previous related works. For that, in this paper, we tried to implement our generative adversarial network with two fully connected sequential models, one as a generator and another as a discriminator. Our generator is trainable which gets random data and tries to create fake human faces. On the other hand, our discriminator gets data from the CelebA dataset and tries to detect that the images generated by the generator are fake or real, and gives feedback to the generator. Based on the feedback the generator improves its model and tries to generate more realistic images.

Author 1: Md. Mahiuddin

Author 2: Md. Khaliluzzaman

Author 3: Md. Shahnur Azad Chowdhury

Author 4: Muhammed Nazmul Arefin

Keywords: Generative adversarial network; fake human faces; generator; discriminator; CelebA dataset

Paper 22: Towards Adapting Metamodelling Technique for an Online Social Networks Forensic Investigation (OSNFI) Domain

Abstract: With the ease of use of smart devices, most data is now kept and exchanged in digital forms such as images, diaries, calendars, movies, and so on. Digital forensic investigation is a new technology that emerged from criminals' who extensively use computers and digital storage devices to commit different types of crimes. To address this issue, a new domain called Online Social Networks Forensic (OSNF) was created to investigate these dynamic crimes perpetrated on social media platforms. OSNFI seeks to obtain, organise, investigate, and visualise user information as direct, objective, and fair evidence. Considering the millions of individuals using social media to share and communicate, they are becoming increasingly relevant for criminal investigations. In forensics investigation of online social network, there are currently major problems such as: lack of structured procedures, the lack of unified automated methods, and the lack of a theoretical context. The use of non-uniform and ad hoc forensic techniques and procedures not only reduces the effectiveness of the process, but also affects the reliability and creditability of the proof in criminal proceedings. As a result, this paper will provide a method derived from the software engineering domain known as metamodelling, which will integrate OSNFI knowledge into an artifact known as a metamodel.

Author 1: Aliyu Musa Bade

Author 2: Siti Hajar Othman

Keywords: Online social networks forensic; online social networks forensic investigation; metamodelling; metamodel; model

Paper 23: Recommendation System based on User Trust and Ratings

Abstract: Recommendation systems aim at providing the user with large information that will be user-friendly. They are techniques based on the individual’s contribution in rating the items. The main principle of recommendation systems is that it is useful for user’s sharing the same interests. Furthermore, collaborative filtering is a widely used technique for creating recommender systems, and it has been successfully applied in many programs. However, collaborative filtering faces multiple issues that affect the recommended accuracy, including data sparsity and cold start, which is caused by the lack of the user's feedback. To address these issues, a new method called “GlotMF” has been suggested to enhance the collaborative filtering method of recommendation accuracy. Trust-based social networks are also used by modelling the user's preferences and using different user's situations. The experimental results based on real data sets show that the proposed method performs better result compared to trust-based recommendation approaches, in terms of prediction accuracy.

Author 1: Mohamed TIMMI

Author 2: Loubna LAAOUINA

Author 3: Adil JEGHAL

Author 4: Said EL GAROUANI

Author 5: Ali YAHYAOUY

Keywords: Recommendation systems; collaborative filtering; trust; social networks

Paper 24: Impervious Surface Prediction in Marrakech City using Artificial Neural Network

Abstract: Determining an impervious surface is one of the most important topics of remote sensing because of its great role in providing information that benefits decision-makers in urban planning, sustainable development goals, and environmental protection. In recent years, a great development in this field has occurred due to the huge improvement in the algorithms and techniques that are used to map impervious surfaces. In this paper, the deep learning technique has been implemented to investigate the extraction of impervious surfaces in Marrakesh city, based on Landsat images. 9000 polygons and 13840 points have been taken to prepare label data by random forest in Google Earth Engine (GEE). In addition, all preprocessing steps for remote sensing images have been implemented in GEE. An artificial neural network (ANN) has been used to determine impervious surfaces. After training and testing the proposed network on Landsat image datasets, precision, accuracy, recall, and F1-score matrix scores were 0.79, 0.98, 0.87, and 0.82, respectively. The experimental results show that this method is efficient and precise for mapping the impervious surfaces of Marrakesh city.

Author 1: Sulaiman Mahyoub

Author 2: Hassan Rhinane

Author 3: Mehdi Mansour

Author 4: Abdelhamid Fadil

Author 5: Waban Al okaishi

Keywords: Deep-learning; remote sensing; artificial neural network ANN; impervious surface

Paper 25: E-AHP: An Enhanced Analytical Hierarchy Process Algorithm for Priotrizing Large Software Requirements Numbers

Abstract: One of the main activities of software requirements analysis is requirements prioritization. The wrong requirements prioritization is risky as it leads to many software failures. The current requirements prioritization techniques can’t deal with large requirement numbers efficiently, which is considered one of their main issues. Many researchers have agreed that the analytical hierarchy process (AHP) is one of the best prioritization techniques as it produces highly accurate results. AHP has two main problems: scalability and inconsistency. These problems have motivated us to propose an improved version of AHP for software requirements prioritization, namely Enhanced AHP (E-AHP). A performance evaluation has been done for the conventional AHP, E-AHP, and one of the recent algorithms that also try to solve the AHP scalability problems, namely removing eigenvalues and introducing the dynamic consistency checking algorithm into AHP (ReDCCahp) algorithms The evaluation shows which algorithm takes the least time, uses the least memory, produces the most consistent and accurate results, and has the highest scalability. The three algorithms have been evaluated by running their codes using different numbers of requirements ranging from 10 to 500. The results show that E-AHP is more scalable, takes the least time, uses the least memory, and produces the most consistent and accurate results compared to the other two algorithms. That becomes remarkable when the number of requirements increases. Therefore, E-AHP is suitable to be applied in large software projects, as it can deal efficiently with the large software requirements numbers.

Author 1: Nahla Mohamed

Author 2: Sherif Mazen

Author 3: Waleed Helmy

Keywords: Requirements engineering; analytical hierarchy process; software engineering; requirements prioritization techniques

Paper 26: Effectiveness of Human-in-the-Loop Sentiment Polarization with Few Corrected Labels

Abstract: In this work, we investigated the effectiveness of adopting Human-in-the-Loop (HITL) aimed to correct automatically generated labels from existing scoring models, e.g. SentiWordNet and Vader to enhance prediction accuracy. Recently, many proposals showed a trend in utilizing these models to label data by assuming that the labels produced are near to ground truth. However, none investigated the correctness of this notion. Therefore, this paper fills this gap. Bad labels result in bad predictions, hence hypothetically, by positioning a human in the computing loop to correct inaccurate labels accuracy performance can be improved. As it is infeasible to expect a human to correct a multitude of labels, we set out to answer the questions of “What is the smallest percentage of corrected labels needed to improve prediction quality against a baseline?” and “Would randomly selecting automatic labels for correction produce better prediction than specifically choosing labels with distinct data points?”. Naïve Bayes (NB) and Decision Tree (DT) were employed on AirBnB and Vaccines public datasets. We could conclude from our results that not all ML algorithms are suited to be used in a HITL environment. NB fared better than DT at producing improved accuracy with small percentages of corrected labels, as low as 1%, exceeding the baseline. When selected for human correction, labels with distinct data points assisted in enhancing the accuracy better than random selection for NB across both datasets, yet partially for DT.

Author 1: Ruhaila Maskat

Author 2: Nurzety Aqtar Ahmad Azuan

Author 3: Siti Auni Amaram

Author 4: Nur Hayatin

Keywords: Human-in-the-loop; few labels; sentiment polarization

Paper 27: Fuzzy Clustering Analysis of Power Incomplete Data based on Improved IVAEGAN Model

Abstract: The scale of data generated by the complex and huge power system during operation is also very large. With the data acquisition of various information systems, it is easy to form the situation of incomplete power data information, which cannot guarantee the efficiency and quality of work, and reduce the security and reliability of the entire power grid. When incomplete data and incomplete data sets are caused by data storage failure or data acquisition errors, fuzzy clustering of data will face great difficulties. The fuzzy clustering of incomplete data of the power equipment is divided into the processing of incomplete data and the clustering analysis of "recovered" complete data. This paper proposes an IVAEGAN-IFCM interval fuzzy clustering algorithm, which uses interval data sets to fill in the incomplete data, and then completes the clustering of interval data. At the same time, the whole numerical data set is transformed into a complete interval data set. The final clustering result is obtained by interval fuzzy mean clustering analysis of the whole interval data set. Finally, the algorithm proposed in this paper and other machine learning training data sets is made for experimental analysis. The experimental results show that the algorithm proposed in this paper can complete incomplete data sets with high precision clustering. Compared with other contrast methods, it shows higher clustering accuracy. Compared with the numerical clustering algorithm, the clustering accuracy is improved by more than 4.3%, and it has better robustness. It also shows better generalization on the artificial data sets and other complex data sets. It is helpful to improve the technical level of the existing power grid and has important theoretical research value and engineering practice significance.

Author 1: Yutian Hong

Author 2: Jun Lin

Keywords: Power system; power equipment; incomplete data; fuzzy clustering; mining algorithm

Paper 28: MCBRank Method to Improve Software Requirements Prioritization

Abstract: Software Requirements Prioritization is an important issue that has a more profound effect on the overall quality of software development. Application of software requirements prioritization provides benefits to minimize risks in software development so that the most important and most impactful requirements are given priority. This paper presents a proposed software requirements prioritization method named MCBRank, which incorporates renowned MoSCoW Method and Case-Based Ranking to improve prioritization correctness. It elaborates on the implementation of MCBRank in an empirical investigation to determine software requirements prioritization for a potential e-library system. The investigation allows the software requirements prioritization process to be implemented by using the proposed MCBRank method. A role-playing empirical investigation with 30 respondents prioritized 31 software requirements, and the results were measured by Cohen Kappa. The kappa results show that MCBRank achieves a better agreement towards the Gold Standard with kappa value of 0.60. Therefore, the investigation results support that MCBRank improves the importance of ranking correctness, representing the stakeholders’ wants and the organization's actual needs for the potential e-library system.

Author 1: Sabrina Ahmad

Author 2: Riftika Rizawanti

Author 3: Terry Woodings

Author 4: Intan Ermahani A. Jalil

Keywords: Requirements prioritization; requirements engineering; software engineering; empirical software engineering

Paper 29: Breast Cancer Detection and Classification using Deep Learning Xception Algorithm

Abstract: Breast Cancer (BC) is one of the leading cause of deaths worldwide. Approximately 10 million people pass away internationally from breast cancer in the year 2020. Breast Cancer is a fatal disease and very popular among women globally. It is ranked fourth among the fatal diseases of different cancers, for example colorectal cancer, cervical cancer, and brain tumors. Furthermore, the number of new cases of breast cancer is anticipated to upsurge by 70% in the next twenty years. Consequently, early detection and precise diagnosis of breast cancer plays an essential part in enhancing the diagnosis and improving the breast cancer survival rate of patients from 30 to 50%. Through the advances of technology in healthcare, deep learning takes a significant role in handling and inspecting a great number of X-ray, Magnetic Resonance Imaging (MRI), computed tomography (CT) images. The aim of this study is to propose a deep learning model to detect and classify breast cancers. Breast cancers has eight classes of cancers: benign adenosis, benign fibroadenoma, benign phyllodes tumor, benign tubular adenoma, malignant ductal carcinoma, malignant lobular carcinoma, malignant mucinous carcinoma, and malignant papillary carcinoma. The dataset was collected from Kaggle depository for breast cancer detection and classification. The measurement that was used in the evaluation of the proposed model includes: F1-score, recall, precision, accuracy. The proposed model was trained, validated and tested using the preprocessed dataset. The results showed that Precision was (97.60%), Recall (97.60%) and F1-Score (97.58%). This indicates that deep learning models are suitable for detecting and classifying breast cancers precisely.

Author 1: Basem S. Abunasser

Author 2: Mohammed Rasheed J. AL-Hiealy

Author 3: Ihab S. Zaqout

Author 4: Samy S. Abu-Naser

Keywords: Breast cancer; deep learning; xception

Paper 30: Improved POS Tagging Model for Malay Twitter Data based on Machine Learning Algorithm

Abstract: Twitter is a popular social media platform in Malaysia that allows for 280-character microblogging. Almost everything that happens in a single day is tweeted by users. Because of the popularity of Twitter, most Malaysians use it daily, providing researchers and developers with a wealth of data on Malaysian users. This paper explains why and how this study chose to create a new Malay Twitter corpus, Malay Part-of-Speech (POS) tags, and a Malay POS tagger model. The goal of this paper is to improve existing Malay POS tags so that they are more compatible with the newly created Malay Twitter corpus, as well as to build a POS tagging model specifically tailored for Malay Twitter data using various machine learning algorithms. For instance, Support Vector Machine (SVM), Naïve Bayes (NB), Decision Tree (DT), and K-Nearest Neighbor (KNN) classifiers. This study’s data was gathered by using Twitter's Advanced Search function and relevant and related keywords associated with informal Malay. The data was fed into machine learning algorithms after several stages of processing to serve as the training and testing corpus. The evaluation and analysis of the developed Malay POS tagger model show that the SVM classifier, as well as the newly proposed Malay POS tags, is the best machine learning algorithm for Malay Twitter data. Furthermore, the prediction accuracy and POS tagging results show that this research outperformed a comparable previous study, indicating that the Malay POS tagger model and its POS were successfully improved.

Author 1: Siti Noor Allia Noor Ariffin

Author 2: Sabrina Tiun

Keywords: Informal Malay; Malay Twitter corpus; Malay POS tagging; Malay POS tagger model; Malay social media texts; Malay POS machine learning

Paper 31: Deep Learning Framework for Locating Physical Internet Hubs using Latitude and Longitude Classification

Abstract: This article proposes framework for determining the optimal or near optimal locations of physical internet hubs using data mining and deep learning algorithms. The framework extracts latitude and longitude coordinates from various data types as data acquisition phase. These coordinates has been extracted from RIFD, online maps, GPS, and GSM data. These coordinates has been class labeled according to decision maker’s preferences using k-mean, density based algorithm (DB Scan and hierarchical clustering analysis algorithms. The proposed algorithm uses haversine distance matrix to calculate the distance between each coordinates rather than the Euclidian distance matrix. The haversine matrix provides more accurate distance surface of a sphere. The framework uses the class labeled data after the clustering phase as input for the classification phase. The classification has been performed using decision tree, random forest, Bayesian, gradient decent, neural network, convolutional neural network and recurrent neural network. The classified coordinates has been evaluated for each algorithms. It has been found that CNN, RNN outperformed the other classification algorithms with accuracy 97.6% and 97.9% respectively.

Author 1: El-Sayed Orabi Helmi

Author 2: Osama Emam

Author 3: Mohamed Abdel-Salam

Keywords: Physical internet hubs (π hubs); deep learning; convolutional neural network (CNN); recurrent neural network (RNN); latitude and Longitude classification

Paper 32: Development of Adaptive Line Tracking Breakpoint Detection Algorithm for Room Sensing using LiDAR Sensor

Abstract: This research focuses on the use of Light Detection and Ranging (LiDAR) sensors for robot localization. One of the most essential algorithms in LiDAR localization is the breakpoint detector algorithm which is used to determine the corner of the room. The previously developed breakpoint detection methods have weaknesses, such as the Adaptive Breakpoint Detector (ABD), could generate dynamic threshold values. The ABD results, on the other hand, still require Line Extraction to obtain the corner breakpoint. Line Extraction method, e.g. Iterative End Point Fit (IEPF), is used to categorize data, resulting in the generation of a line pattern as an interpretation of a wall. The computational method for obtaining the corner breakpoint becomes longer as the line is extracted. To address this issue, our algorithm proposes a new threshold area in the form of an ellipse with the threshold value parameter obtained from previously identified room size and sensor characteristics. As a result the corner breakpoint detection becomes more adaptive. The goal of this research is to create an Adaptive Line Tracking Breakpoint Detector (ALTBD) approach that will reduce the computing time required to detect corner breakpoints. Furthermore, the Line Extraction method required for corner breakpoint detection is modified in the ALTBD. To distinguish between the edge of the wall and the corner of the room, the boundary value is increased. The ALTBD method was tested in a simulation arena comprised of multiple rooms and halls. According to the results, the ALTBD computation time is faster in detecting corner breakpoints than the ABD IEPF method, also the accuracy for determining the position of the robot was improved.

Author 1: Deddy El Amin

Author 2: Karlisa Priandana

Author 3: Medria Kusuma Dewi Hardhienata

Keywords: LiDAR; breakpoint detector; robot localization; corner detection; line segmentation

Paper 33: Alignment of Software System Level with Business Process Level: Resolving Syntactic and Semantic Conflicts

Abstract: Information systems help organizations manage their entities with innovative technologies. These entities are often very different in nature. In this paper, we consider a business process level based on a set of Business Process Model and Notation (BPMN) models and a software system level based on a Unified Modeling Language (UML) class diagram. The differences between these entities make them difficult to align. In addition, an organization’s BPMN models may be designed by different teams, which can cause syntactic and semantic heterogeneities. We present the first step of our proposed approach for aligning a software system level with a business process level without conflict (redundancy and lost information). Syntactic and semantic rules based on ontologies and other resources for comparing BPMN models are described, as well as a process for transforming BPMN models into UML model.

Author 1: Samia Benabdellah Chaouni

Author 2: Maryam Habba

Author 3: Mounia Fredj

Keywords: Information system alignment; business process; software system; Business Process Model and Notation (BPMN); Unified Modelling Language (UML); class diagram; ontology; semantic aspects

Paper 34: Multi-Task Reinforcement Meta-Learning in Neural Networks

Abstract: Artificial Neural Networks (ANN) is one of the main and widespread tools for creating intelligent systems. And, they are actively used for data analysis in many areas such as robotics, computer vision, natural language processing, etc. The learning process of ANN is one of the most labor-intensive stages in ANN. There are many different modifications of ANNs and methods for their training. Currently, deep neural networks are becoming one of the most popular methods of machine learning due to their effectiveness in areas such as speech recognition, medical informatics, computer vision, etc. It is known that ANN training depends on the type of input data. In this paper, reinforcement learning is considered, as popular method used in cases where information is reinforced by signals from the external environment with which the model interacts. The purpose of this paper is to develop a reinforcement meta-learning algorithm that would be efficient in terms of quality and speed of learning. However, despite the significant scientific progress in deep learning, existing algorithms are not efficient enough to solve problems in the real world. In addition, such algorithms require a significant amount of learning time, which complicates the development process. To solve these problems, the use of meta-learning or “learning to learn” algorithms has recently been especially relevant. The paper proposes an approach to reinforcement meta-learning using a multitasking weight optimizer. experimentally shown that the proposed approach is more efficient than the known MAML (Model-Agnostic Meta-Learning) algorithm. The proposed MAML SPSA-Track method shows an improvement in efficiency by an average of 4%, and MAML SPSA-Delta by 8%, respectively. Moreover, the last algorithm spends on average 2 times less time on push-v2 and pick-place-v2 tasks.

Author 1: Ghazi Shakah

Keywords: Multitasking; meta-learning; reinforcement learning; neural networks; optimization

Paper 35: Data Fusion Model of Road Sensors based IoT Feature Clustering

Abstract: The collection of traffic data can play a role in analyzing and predicting highway design, planning, and real-time traffic management. The accuracy requirements for road dynamic data collection are low, and the accuracy is usually 3%-5%. However, it is required that vehicles can pass at high speed and obtain traffic information such as vehicle classification and vehicle speed. The prerequisite for the application of Internet of Things (IoT) technology to road information monitoring lies in the research and development of sensor technology in the perception layer and communication technology in the network layer, so that can obtain a large amount of perception data to serve the development and application of algorithms. To achieve the goal of low-cost and long-term monitoring of comprehensive traffic information and road service status information, this paper constructs a road vibration monitoring system, carries out road vibration monitoring under complex road environments, and proposes a traffic information monitoring method driven by road vibration data. By deploying the pavement vibration monitoring system in the actual road, the original signal of pavement vibration under the action of vehicle moving load is obtained. Through smooth processing and eigenvalue extraction, the monitoring of vehicle speed, wheelbase driving direction, vehicle load position and traffic flow is realized. The experimental results prove that the analysis of the road dynamic response under working conditions, as well as smoothing processing and eigenvalue extraction, the numerical modeling method in this paper realizes the monitoring of the position of the vehicle load and the traffic flow. The calculation error of vehicle speed and wheelbase is within ±4%, which is helpful to find the characteristic index of road vibration signal for evaluating road service status, and provides a reference for the application of road vibration response in road damage early warning and scientific maintenance.

Author 1: Hua Yang

Keywords: Traffic data; Internet of Things (IoT) technology; perception sensors; vibration monitoring; k-means++ algorithm

Paper 36: Automatic State Recognition of Multi-type Protection Platens based on YOLOv5 and Parallel Multi-Decision Algorithm

Abstract: A protection platen is a vital component in relay protection systems. The manual inspection of protection platen states is long-term repetitive work with low efficiency and imposes a heavy burden on workers. In this work, we propose a new system to automatically detect the states of multi-type protection platens in images. This system can classify two protection platen categories and further recognize the states of protection platens. For the classification of protection platen types, we propose a new algorithm that automatically detects two protection platen types based on HSV (Hue, Saturation, Value) color space weighting operators. The proposed operators quantify the color variation in the protection platen and reduce the influence of environmental factors. With respect to the state recognition of protection platens, the Type-I protection platen states are automatically classified by the YOLOv5 (You only look once version 5) network. Since the Type-II protection platen has three primary states and more complicated structures, we investigate a new parallel multi-decision algorithm to recognize the states of Type-II protection platens based on the newly proposed watershed-color space difference-shape feature (W-CD-SF) method and the YOLOv5 network. The W-CD-SF technique can segment the protection platens and extract their shape features automatically. This multi-decision mechanism improves the robustness and generalization of state recognition. Experiments were conducted on the collected protection platen images containing 8,969 protection platens. The recognition accuracies of protection platen states exceed 95%. This system can provide auxiliary detection and long-term monitoring of protection platen states.

Author 1: Ying Zhang

Author 2: Yihui Zhang

Author 3: Hao Wu

Author 4: Boxiong Fu

Author 5: Ling He

Keywords: Protection platen; parallel multi-decision; YOLOv5 network; watershed-color space difference-shape feature

Paper 37: Dual Authentication for Payment Request Verification Over Cloud using Bilinear Dual Authentication Payments Transaction Protocol

Abstract: There has been a recent explosion in the number of mobile network payment gateways that enable users to access services through a variety of devices. Mobile payment gateway security is complicated by a number of difficult-to-solve issues. As digital technology has progressed over the last decade, mobile payment mechanisms have gained a lot of interest. In the internet industry, these standards might have a significant impact on service quality. However, the most important aspect to consider when using these systems is their accountability, which assures confidence between the parties engaged in the financial transactions. Mobile payments may be easy, quick and secure. On the other hand, they may be rather pricey and are still susceptible to problems caused by technology. Specifically, mobile payments won't be able to go through at all if there are any problems with the host phone. For this reason, in this article a mobile payment mechanism that uses secure bilinear dual authentication. Using Bullet hash Maximum distance separable (MDS) and the mutate Hellman algorithm, our payment protocol incorporates all of the essential security characteristics to establish confidence between the parties. To put it another way, accountability is assured by mutual authentication and non-repudiation. Conflicts that may emerge in the course of a payment transaction may be resolved using our strategy. Scyther is used to test our suggested protocol's empirical performance.

Author 1: A. Saranya

Author 2: R. Naresh

Keywords: Mobile payment; transaction protocol; bullet hash maximum distance separable; mutate Hellman algorithm

Paper 38: An Efficient Approach towards Vehicle Number Estimation with Ad-hoc Network under Vehicular Environment

Abstract: Ad-hoc network usability extends the application for Dedicated Short-Range Communication. This type of ad-hoc network technology is non infrastructure and due to this fact, it can be used for Direct Short Range Communication System to provide the quick and real time message to the vehicular operator to prevent the damage and fatality of life by meeting accidents and crashes. In this paper, we present a holistic approach to estimate the number of vehicles in specified range of one KM distance. The designed system for vehicle number estimation is based on the Time Division Multiple Access mechanism which further estimates the number of reserved slots by vehicular nodes. This estimation methodology is tested under the digital simulator and approximately 34 number of vehicles for 24 seconds are defined to test the slot reservation. We found that in case of vehicular nodes greater than 20, slot reservation accuracy is 95% and when the vehicular nodes are less than 20 then the slot reservation is 100%.

Author 1: Yuva Siddhartha Boyapati

Author 2: Shallaja Salagrama

Author 3: Vimal Bibhu

Keywords: Vehicular ad hoc networking; hidden vehicle; visible vehicle; time division multiple access; dedicated short range communication

Paper 39: Virtual Tourism and Digital Heritage: An Analysis of VR/AR Technologies and Applications

Abstract: During the time of the pandemic, travel restrictions have impacted the tourism industry with an estimated loss of more than a trillion USD; however, at the same time, we have seen a significant increase in profits for the industries which empower remote connectivity. Various studies have identified the positive impact of virtual tourism, in which tourists can be attracted by providing a VR/AR-based experience of the destination. Similarly, virtual, mixed, and augmented realities are being used to enhance user experience in digital heritage and its preservation. With emerging technologies and increasing demand for e-tourism (due to travel restrictions), there is a need to review the technological changes and analyze user requirements with respect to virtual tourism. This paper provides a literature review of the latest technologies and applications that can potentially benefit the virtual tourism and digital heritage industry. We also provide an analysis of its impact on user experience, awareness, and interest, as well as the pros and cons of virtual experiences, which may benefit the research community about the current spectrum of virtual tourism and digital heritage.

Author 1: Muhammad Shoaib Siddiqui

Author 2: Toqeer Ali Syed

Author 3: Adnan Nadeem

Author 4: Waqas Nawaz

Author 5: Ahmad Alkhodre

Keywords: Virtual tourism; digital heritage; virtual reality; user experience

Paper 40: Using a Fuzzy-Bayesian Approach for Predictive Analysis of Delivery Delay Risk

Abstract: Although one of the major roles of delivery logistics activities is to ensure a good quality of customer service, certain risks such as damage, delay, return of transported goods occur quite often. This makes risk control and prevention one of the requirements of transport supply chain quality. The article focuses on the analysis of the risk of delay, which is often considered fundamental for the quality of service and as a center of additional costs related to the violation of time windows. Such a risk can harm the image of a supplier, which can even lead to the loss of customers in case of recurrence. The aim of the following case study is the development of a fuzzy-bayesian approach that anticipates, by predictive analysis combining Bayesian networks (BNs) and Fuzzy logic, the possible delays affecting the smooth running of a delivery operation. The results of the implementation of the late delivery risk prediction model are validated by verifying three axioms. In addition, a sensitivity and scenario analysis is performed to validate the model and identify the parameters that have the most adverse impact on the occurrence of such a risk. These results can help carriers/transport providers to minimize potential late deliveries. In addition, the developed model can be used as a basis for different types of predictions in the field of freight transport as well as in other research areas.

Author 1: Ouafae EL Bouhadi

Author 2: Monir Azmani

Author 3: Abdellah Azmani

Author 4: Mouna Atik el ftouh

Keywords: Delivery logistics; risk management; predictive analysis; bayesian network; fuzzy logic

Paper 41: Mining Educational Data to Analyze the Student’s Performance in TOEFL iBT Reading, Listening and Writing Scores

Abstract: Student scores in TOEFL IBT reading, listening, and writing may reveal weaknesses and deficiencies in educational institutions. Traditional approaches and evaluations are unable to disclose the significant information hidden inside the student's TOEFL score. As a result, data mining approaches are widely used in a wide range of fields, particularly education, where it is recognized as Educational Data Mining (EDM). Educational data mining is a prototype for handling research issues in student data which can be used to investigate previously undetected relationships in a huge database of students. This study used the EDM to define the numerous factors that influence students' achievement and to create observations using advanced algorithms. The present study explored the relationship among university students’ previous academic experience, gender, student place and their current course attendance within a sample of 473 (225 male and 248 female). Educational specialists must find out the causes of student dropout in TOEFL scores. The results of the study showed that the model could be suitable for investigation of important aspects of student outcomes, the present research was supposed to use the statistical package for social sciences (SPSS V26) for both descriptive and inferential statistics and multiple linear regressions to improve their scores.

Author 1: Khaled M. Hassan

Author 2: Mohammed Helmy Khafagy

Author 3: Mostafa Thabet

Keywords: Educational data mining; students score; linear regression; TOEFL; Statistics

Paper 42: Average Delay-based early Congestion Detection in Named Data of Health Things

Abstract: The Internet of Health Things (IoHT) is receiving more attention from researchers because of its wide use in the healthcare field. IoHT refers to medical devices whose main purpose is to transmit health data in a secure and lossless manner between them and healthcare personnel. However, in a medical emergency, sensors transmit vital patient data simultaneously and frequently, increasing the risk of congestion and packet loss. This problem is highly undesirable in an IoHT system, leading to undesirable results. To address this issue, a new approach based on Named Data Networking (NDN) (which is considered as the most appropriate internet architecture for IoT systems) is proposed to control congestion in IoHT systems. The proposed approach, Average delay-based early congestion Detection (ADCD), detects and controls congestion at consumer nodes by calculating the average queuing delay based on the one-way delay similar to that proposed in Sync-TCP. Then according to the calculated value, ADCD divides the network into three states: no-congested state, less congested state, and heavily congested state. The adjustment of the congestion window size is done according to the state of the network. ADCD was implemented in ndnSIM and compared to the Interest Control Protocol ICP. The results show that ADCD maximizes bandwidth utilization compared to ICP and maintains a reasonable delay.

Author 1: Asmaa EL-BAKKOUCHI

Author 2: Mohammed EL GHAZI

Author 3: Anas BOUAYAD

Author 4: Mohammed FATTAH

Author 5: Moulhime EL BEKKALI

Keywords: Named data networking; internet of health things; congestion control; congestion detection

Paper 43: Novel Approach for Spatiotemporal Weather Data Analysis

Abstract: Massive volumes of multidimensional array-based spatiotemporal data are generated by climate observations and model simulations. The growth in climate data leads to new opportunities for climate studies at multiple spatial and temporal scales. Managing, analyzing and processing of big climate data is considered to be challenging because of huge data volume. In this work multidimensional climate data such as precipitation and temperature are processed and analyzed in the Spark MapReduce Platform, since Spark platform is computationally more efficient than Hadoop-MapReduce Platform of same configuration. In temporal scale monthly and seasonal analysis of climate data has been carried out to understand the regional climate. The prediction of Rainfall on monthly and seasonal time scales is very much important for planning and devising agricultural strategies and disaster management, etc. As the prediction of climate state is very challenging, in this study an attempt is being made for the prediction of the rainfall using the time series analysis in the same framework. As a test case the time series approach such as Auto Regression Integrated Moving Average (ARIMA) has been used for the prediction of rainfall. The proposed approach is evaluated and found to be accurate in the analysis and prediction of climate data and this will surely guide for the researcher for better understanding of the climate and its application to multiple sectors.

Author 1: Radhika T V

Author 2: K C Gouda

Author 3: S Sathish Kumar

Keywords: Spatiotemporal; big climate data; spark; ARIMA

Paper 44: Prediction Models to Effectively Detect Malware Patterns in the IoT Systems

Abstract: The Widespread use of the Internet of Things (IoT) has influenced many domains including smart cities, cameras, wearables, smart industrial equipment, and other aspects of our daily lives. On the other hand, the IoT environment deals with a massive volume of data that needs to be kept secure from tampering or theft. Detection of security attacks against IoT context requires intelligent techniques rather than relying on signature matching. Machine learning (ML) and Deep Learning (DL) approaches are efficient to detect these attacks and predicting intrusion behavior based on unknown patterns. This study proposes the application of five deep and ML techniques for identifying malware in network traffic based on the IoT-23 dataset. Random Forest, Catboost, XGBoost, Convolutional Neural Network, and Long Short-Term Memory (LSTM) models are among the classifiers utilized. These algorithms have been selected to provide lightweight security systems to be deployed in the IoT devices rather than a centralized approach. The dataset was preprocessed to remove unnecessary or missing data, and then the most significant features were extracted using a feature engineering technique. The highest overall accuracy achieved was 96% by applying all classifiers except LSTM which recorded a lower accuracy.

Author 1: Rawabi Nazal Alhamad

Author 2: Faeiz M. Alserhani

Keywords: Internet of Things (IoT); malware deletion; random forest; Catboost; convolutional neural network; long short-term memory (LSTM); XGBoost

Paper 45: Power Grid Resource Integration of Enterprise Middle Station based on Analytic Hierarchy Process

Abstract: With the transformation from smart grid to power Internet of Things, new power businesses such as power grid automation and power quality monitoring are constantly emerging. The load environment of power grid is changeable. In order to meet the needs of multi-service, the integrated access scheme for power grid resources in power enterprises is gradually diversified, which brings challenges to the unified management and control of power grid communication network. In this paper, SDN technology is used to improve the operation and maintenance management and control of power communication network, which aims at the integration scheme of power grid resources in power enterprises. Based on the controller cluster technology, combined with the new power business requirements, this paper designs a software-defined network centralized control architecture of the new business of power communication network. The architecture realizes the operation and maintenance management of network resources under the centralized control architecture of typical enterprise scenarios, such as power grid enterprises. The convergence speed is improved by 27%. The minimum value of iterative convergence is 31% better than that of other methods. The system requirement is reduced by 13.5%, which is helpful to improve the efficiency of node dynamic allocation and ensure the need of large-capacity data transmission of smart grid. The research in this paper can realize the two-way interaction, real-time expansion and unified deployment of power business in the future, and promote the intensive and lean development of power communication network.

Author 1: Shaobo Liu

Author 2: Li Chen

Author 3: Wenyuan Bai

Author 4: Zhen Zhang

Author 5: Fupeng Li

Keywords: Power grid enterprises; SDN; power grid resources; centralized control architecture; power communication network

Paper 46: A Screening System for COVID-19 Severity using Machine Learning

Abstract: COVID-19 disease can be classified into various stages depending on the severity of the patient. Patients in severe stages of COVID-19 need immediate treatment and should be placed in a medical-ready environment because they are at high risk of death. Thus, hospitals need a fast and efficient method to screen large numbers of patients. The enormous amount of medical data in public repositories allows researchers to gain information and predict possible outcomes. In this study, we use a publicly available dataset from Springer Nature repository to discuss the performance of three machine learning techniques for prediction of severity of COVID-19: Random Forest (RF), Naïve Bayes (NB) and Gradient Boosting (GB). These techniques were selected for their good performance in medical predictive analytics. We measured the performance of the machine learning techniques using six measurements (accuracy, precision, recall, F1-score, sensitivity and specificity) in predicting COVID-19 severity. We found that RF generates the highest performance score, which is 78.4, compared with NB and GB. We also conducted experiments with RF to establish the critical symptoms in predicting COVID-19 severity, and the findings suggested that seven symptoms are substantial. Overall, the performance of various machine learning techniques to predict severity of COVID-19 using electronic health records indicates that machine learning can be successfully applied to determine specific treatment and effective triage.

Author 1: Abang Mohd Irham Amiruddin Yusuf

Author 2: Marshima Mohd Rosli

Author 3: Nor Shahida Mohamad Yusop

Keywords: Severity prediction; COVID-19; random forest; Naïve Bayes; gradient boosting

Paper 47: Electronic Personal Health Record Assessment Methodology: A Review

Abstract: ePHR (Electronic Personal Health Record) is not a new concept in the era of electronic health information. The advantages of ePHR in improving health outcomes through patient empowerment have been recognized globally and almost all countries that implement electronic health records (EHR) have created ePHR. This study identifies the components of the ePHR implementation study methodology that has been conducted throughout the country. The types of ePHR studies selected were adoption studies, acceptance studies, readiness studies, and evaluation studies. This study’s systematic literature review process is identification, screening, eligibility, data abstraction, and analysis. A total of 16 final journals were analyzed from 173 journals identified from 5 databases (Science Direct, WoS, Scopus, JMIR, and PubMed) regardless of the year of publication until April 1st, 2021. Among the findings based on the four objectives of the study, there are two findings that are considered important and interesting by the author; first, the existence of 22 additional variables to the evaluation model by almost all studies in this study which shows a clear need to improve the evaluation model which is the TAM Model. Second, although the proposal of conducting a scientific study to evaluate the perspective of ePHR stakeholders before ePHR is developed only appeared once, based on this study and the knowledge of the authors, it is a starting point for the successful implementation of ePHR. These two findings contribute to the recommendations for the best design of the ePHR implementation study described in this paper.

Author 1: Dirayana Kamarudin

Author 2: Nurhizam Safie

Author 3: Hasimi Sallehudin

Keywords: Personal health record; ePHR; ePHR evaluation variable

Paper 48: Contribution of Experience in the Acquisition of Physical Sciences Case of High School Students Qualifying

Abstract: This work deals with the importance of experimental practice in the teaching of physical sciences. By practice, we rarely mean practical work carried out in the laboratory. We examined the relationship between students' knowledge of physical science and how practice may or may not help them understand chemical and/or physical concepts. What emerges from a survey distributed to our students is that they are very favorable of the use of the practice. The problem posed in this work consists in identifying the impact of experiments on the acquisition of knowledge and in responding to the problems of learning through experience in the short and medium term. The analysis of the answers allowed us to conclude that the experiment in class, by the teacher, helps to understand the physical and chemical phenomena and can be done before or after the study of the theory. The length and difficulty of practical work sometimes worry students, trying to follow the protocol step by step. This fact underscores the importance of clarity of purpose, through which students can be guided toward questioning what is expected of them, such as knowing how their knowledge has increased.

Author 1: Zineb Azar

Author 2: Oussama Dardary

Author 3: Jabran Daaif

Author 4: Malika Tridane

Author 5: Said Benmokhtar

Author 6: Said Belaaouad

Keywords: Experimental practice; physical sciences; students; practical Work PW

Paper 49: Fit-Gap Analysis: Pre-Fit-Gap Analysis Recommendations and Decision Support Model

Abstract: Enterprise Resource Planning (ERP) system has been defined as a configurable Commercial Off-The-Shelf (COTS) system integrated into multiple business functions. For most companies, adopting ERP has become necessary to maintain market competitiveness. However, ERP implementation is still critical because project success depends on multiple parameters and involves several stakeholders. This article deals with the Fit-Gap analysis stage, which is an essential step in ERP implementation. This study was carried out through a literature review and interviews with experts to gather information and support stakeholders toward a successful Fit-Gap phase. It presents a set of recommendations for clients and consultants to consider before starting the Fit-Gap Analysis phase, and it presents an approach, with a decision support model represented as Business Process Modelling Notation (BPMN) based on several parameters to be used during the Fit-Gap Analysis stage to bridge gaps. The results obtained are intended for clients and consultants to make the most rational decision to bridge gaps based on the recommendations found, the approach and the decision support models presented.

Author 1: LAHLOU Imane

Author 2: MOTAKI Nourredine

Author 3: SARSRI Driss

Author 4: L’YARFI Hanane

Keywords: ERP systems; misfit; customisation; fit-gap analysis

Paper 50: An Improved Deep Learning Model of Chili Disease Recognition with Small Dataset

Abstract: Due to its tasty and spicy fruit with nutritional qualities, chili is a demanding crop widely farmed around the world. Hence, it is essential to accurately determine the health status of chili for agricultural productivity. Recent years have seen impressive results in recognition fields due to deep learning approaches. However, deep learning models’ networks need an abundant data to perform well and collecting enormous data for the networks is time-consuming and resource-intensive. A data augmentation method is proposed to overcome this problem. It was applied to a small dataset of healthy and diseased chili leaf by utilizing geometric transformation method. Eventually, two deep learning models of CNN and ResNet-18 were evaluated using augmented and original datasets. From a series of experiment, it can be concluded that the trained deep learning models using original and augmented datasets perform better with an average accuracy performance of 97%.

Author 1: Nuramin Fitri Aminuddin

Author 2: Zarina Tukiran

Author 3: Ariffuddin Joret

Author 4: Razali Tomari

Author 5: Marlia Morsin

Keywords: Chili leaf; deep learning; data augmentation; geometric transformation

Paper 51: Arabic Image Captioning: The Effect of Text Pre-processing on the Attention Weights and the BLEU-N Scores

Abstract: Image captioning using deep neural networks has recently gained increasing attention, mostly for English langue, with only few studies in other languages. Good image captioning model is required to automatically generate sensible, syntactically and semantically correct captions, which in turn requires good models for both computer vision and natural language processing. The process is more challenging in case of data scarcity, and languages with complex morphological structures like the Arabic language. This was the reason why only limited number of studies have been published for Arabic image captioning, compared to those of English language. In this paper, an efficient deep learning model for Arabic image captioning has been proposed. In addition, the effect of using different text pre-processing methods on the obtained BLEU-N scores and the quality of generated images, as well as the attention mechanism behavior were investigated. Furthermore, the “THUMB” framework to assess the quality of the generated captions is used -for the first time- for Arabic captions’ evaluation. As shown in the results, a BLEU-4 score of 27.12, has been achieved, which is the highest obtained results so far, for Arabic image captioning. In addition, the best THUMB scores were obtained, compared to previously published results on common images.

Author 1: Moaz T. Lasheen

Author 2: Nahla H. Barakat

Keywords: Arabic image captioning; computer vision; deep learning; image captioning; natural language processing

Paper 52: Identification of Human Sperm based on Morphology Using the You Only Look Once Version 4 Algorithm

Abstract: Infertility is a crucial reproductive problem experienced by both men and women. Infertility is the inability to get pregnant within one year of sexual intercourse. This study focuses on infertility in men. Many causes that can cause infertility in men including sperm quality. Currently, identification of human sperm is still done manually by observing the sperm with the help of humans through a microscope, so it requires time and high costs. Therefore, high technology is needed to determine sperm quality in the form of deep learning technology based on video. Deep learning algorithms support this research in identifying human sperm cells. So deep learning can help detect sperm video automatically in the process of evaluating sperm cells to determine infertility. We use deep learning technology to identify sperm using the You Only Look Once version 4 (YOLOv4) algorithm. Purpose of this study was to analyze the level of accuracy of the YOLOv4 algorithm. The dataset used is sourced from a VISEM dataset of 85 videos. The results obtained are 90.31% AP (Average Precision) for sperm objects and 68.19% AP (Average Precision) for non-sperm objects, then for the results of the training obtained by the model 79.58% mAP (Mean Average Precision). Our research show result about identification of human sperm using YOLOv4. The results obtained by the YOLOv4 model can identify sperm and non-sperm objects. The output on the YOLOv4 model is able to identify objects in the test data in the form of video and image.

Author 1: Aristoteles

Author 2: Admi Syarif

Author 3: Sutyarso

Author 4: Favorisen R. Lumbanraja

Author 5: Arbi Hidayatullah

Keywords: Classification; deep learning; identification; sperm; sperm head; you only look once version 4

Paper 53: Brain Tumor Segmentation and Classification from MRI Images using Improved FLICM Segmentation and SCA Weight Optimized Wavelet-ELM Model

Abstract: Image segmentation is an essential technique of brain tumor MRI image processing for automated diagnosis of an image by partitioning it into distinct regions referred to as a set of pixels. The classification of the tumor affected and non-tumor becomes an arduous task for radiologists. This paper presents a novel image enhancement based on the SCA (Sine Cosine Algorithm) optimization technique for the improvement of image quality. The improved FLICM (Fuzzy Local Information C Means) segmentation technique is proposed to detect the affected regions of brain tumor from the MRI brain tumor images and reduction of noise from the MRI images by introducing a fuzzy factor to the objective function. The SCA weight-optimized Wavelet-Extreme Learning Machine (SCA-WELM) model is also proposed for the classification of benign tumors and malignant tumors from MRI brain images. In the first instance, the enhanced images are undergone improved FLICM Segmentation. In the second phase, the segmented images are utilized for feature extraction. The GLCM feature extraction technique is considered for feature extraction. The extracted features are aligned as input to the SCA-WELM model for the classification of benign and malignant tumors. The following dataset (Dataset-255) is considered for evaluating the proposed classification approach. An accuracy of 99.12% is achieved by the improved FLICM segmentation technique. The classification performance of the SCA-WELM is measured by sensitivity, specificity, accuracy, and computational time and achieved 0.98, 0.99, 99.21%, and 97.2576 seconds respectively. The comparison results of SVM (Support Vector Machine), ELM, SCA-ELM, and proposed SCA-WELM models are presented to show the robustness of the proposed SCA-WELM classification model.

Author 1: Debendra Kumar Sahoo

Author 2: Satyasis Mishra

Author 3: Mihir Narayan Mohanty

Keywords: Sine cosine algorithm; extreme learning machine; fuzzy c means; GLCM feature; support vector machine

Paper 54: A Machine Learning and Multi-Agent Model to Automate Big Data Analytics in Smart Cities

Abstract: The objective of this paper is to present an architecture to improve the process of automating big data analytics using a multi-agent system and machine learning techniques, to support the processing of real time big data streams and to enhance the process of decision-making for urban planning and management. With the rapidly evolving information technologies, and their utilization in many areas such as smart cities, social networks, urban management and planning, massive data streams are generated and need an efficient approach to deal with. The proposition in this paper adopts the concept of smart data which focuses on the value aspect from big data. The proposed architecture is composed of three layers: data acquisition and storage, data management and processing and the service layer, based on a multi-agent system to automate the big data analytics; the proposed model describe the functionalities of the system and the collaboration between agents, these autonomous entities receive data streams in real time, they perform operations of preprocessing, big data analytics and storage into a Hadoop cluster. The techniques of machine learning are also used to enhance the process of decision making, such the use of classification algorithms to predict habitat type based on the characteristics of a population to help making efficient urban planning decisions. The proposed system can serve as a platform to support data management and to conduct effective decision-making in smart cities.

Author 1: Fouad SASSITE

Author 2: Malika ADDOU

Author 3: Fatimazahra BARRAMOU

Keywords: Big data analytics; machine learning; smart data; multi-agent system; automation; decision-making; urban planning

Paper 55: Brain Tumor Detection using MRI Images and Convolutional Neural Network

Abstract: A brain tumor is the cause of abnormal growth of cells in the brain. Magnetic resonance imaging (MRI) is the most practical method for detecting brain tumors. Through these MRIs, doctors analyze and identify abnormal tissue growth and can confirm whether the brain is affected by a tumor or not. Today, with the emergence of artificial intelligence techniques, the detection of brain tumors is done by applying the techniques and algorithms of machine learning and deep learning. The advantages of the application of these algorithms are the quick prediction of brain tumors, fewer errors, and greater precision, which help in decision-making and in choosing the most appropriate treatment for patients. In the proposed work, a convolution neural network (CNN) is applied with the aim of detecting the presence of a brain tumor and its performance is analyzed. The main purpose of this article is to adopt the approach of convolutional neural networks as a machine learning technique to perform brain tumor detection and classification. Based on training and testing results, the pre-trained architecture model reaches 96% in precision and classification accuracy rates. For the given dataset, CNN proves to be the better technique for predicting the presence of brain tumors.

Author 1: Driss Lamrani

Author 2: Bouchaib Cherradi

Author 3: Oussama El Gannour

Author 4: Mohammed Amine Bouqentar

Author 5: Lhoussain Bahatti

Keywords: Brain tumor; machine learning; convolutional neural network; MRI images

Paper 56: Firefly Algorithm with Mini Batch K-Means Entropy Measure for Clustering Heterogeneous Categorical Timber Data

Abstract: Clustering analysis is the process of identifying similar patterns in various types of data. Heterogeneous categorical data consists of data on ordinal, nominal, binary, and Likert scales. The clustering solution for heterogeneous data clustering remains difficult due to partitioning complex and dissimilarity features. It is necessary to find a solution to high-quality clustering techniques to efficiently determine the significant features of the data. This paper emphasizes using the firefly algorithm to reduce the distance gap between features and improve clustering performance. To obtain an optimal global solution for clustering, we proposed a hybrid of mini-batch k-means (MBK) clustering-based entropy distance measures (EM) with a firefly optimization algorithm (FA). This study compares the performance of hybrid K-Means, Agglomerative, DBSCAN, and Affinity clustering models with EM and FA. The evaluation uses a variety of data from the timber perception survey dataset. In terms of performance, the proposed MBK+EM+FA has superior and most effective clustering. It achieves a higher accuracy of 96.3 percent, a 97 percent F-measure, a 98 percent precision, and a 97 percent recall. Other external assessments revealed that the Homogeneity (HOMO) is 79.14 percent, the Fowlkes-Mallows Index (FMI) is 93.07 percent, the Completeness (COMP) is 78.04 percent, and the V-Measure (VM) is 78.58 percent. Both proposed MBK+EM+FA and MBK+EM took about 0.45s and 0.35s to compute, respectively. The excellent quality of the clustering results does not justify such time constraints. Surprisingly, the proposed model reduced the distance measure of all heterogeneous features. The future model could put heterogeneous categorical data from a different domain to the test.

Author 1: Nurshazwani Muhamad Mahfuz

Author 2: Marina Yusoff

Author 3: Muhammad Shaiful Nordin

Author 4: Zakiah Ahmad

Keywords: Clustering; mini batch k-means; entropy; heterogeneous categorical; firefly optimization algorithm

Paper 57: Improved Spatial Invariance for Vehicle Platoon Application using New Pooling Method in Convolution Neural Network

Abstract: The imbalanced dataset is a prominent concern for automotive deep learning researchers. The proposed work provides a new mixed pooling strategy with enhanced performance for imbalanced vehicle dataset based on Convolution Neural Network (CNN). Pooling is crucial for improving spatial invariance, processing time, and overfitting in CNN architecture. Max and average pooling are often utilized in contemporary research articles. Both techniques of pooling have their own advantages and disadvantages. In this study, the advantages of both pooling algorithms are evaluated for the classification of three vehicles: car, bus, and truck for imbalanced datasets. For each epoch, the performance of max pooling, average pooling, and the new mixed pooling method was assessed using ROC, F1-score, and error rate. Comparing the performance of the max-pooling method to that of the average pooling method, it has been found that the max-pooling method is superior. The performance of the proposed mixed pooling approach is superior to that of the maximum pooling and average pooling methods. In terms of Receiver Operating Characteristics (ROC), the proposed mixed pooling technique is approximately 2 per cent better than the maximum pooling method and 8 per cent better than the mixed pooling method. Using a new pooling technique, the classification performance with an imbalanced dataset is improved, and also a novel mixed pooling method is proposed for the classification of vehicles.

Author 1: M S Sunitha Patel

Author 2: Srinath S

Keywords: Average pooling; convolution neural network; imbalance dataset; max pooling; mixed pooling

Paper 58: Framework to Develop a Resilient and Sustainable Integrated Information System for Health Care Applications: A Review

Abstract: The reconstruction of the health sector amidst the forth industrial revolution has been confronted with many challenges. Many benefits have been attributed to the vital role played by technology in realizing and constructing a robust health information system. However, amidst the digitalization in the healthcare system, several challenges such as integration and fragmentation have been affecting the structure of the Health Information Systems (HIS) which subsequently influences decision making and resource allocation. Therefore, this paper through a comprehensive systematic review afford a proposition for a develop a resilient and sustainable information system for Health Care applications. The study reveals the parallel impact of health information technology application in the healthcare arena and highlight the need for more in-depth research on HIS that incorporate novel scientific methods. Additional this study also presents a body of evident that reveal the inadequacies of the HIS to tackle the constant transformative changes presently confronting the global healthcare systems.

Author 1: Ayogeboh Epizitone

Keywords: Health information system; integrated information system; e-health; bioinformatics

Paper 59: TEC Forecasting using Optimized Variational Mode Decomposition and Elman Neural Networks

Abstract: Forecasting the ionosphere layer’s total electronic content (TEC) is crucial for its impact on satellite signals and global positioning systems (GPS) and the ability to predict earthquakes. The existing statistical-based forecasting models such as ARMA, ARIMA, and HW suffered from the TEC non-stationarity nature, which requires algorithmic handling of the forecasting and the mathematical part. This study proposes a hybrid method that incorporates several components and is designated as Optimized Variational Mode Decomposition with Recursive Neural Network Forecasting (OVMD-RNN) to forecast TEC. Before using the Elman Network to train each component, Variational Mode Decomposition (VMD) was used to decompose the signal into its essential stationary components. In addition, the proposed method includes an optimization algorithm for determining the best VMD decomposer parameters. The GPS Ionospheric Scintillation and TEC Monitor (GISTM) at Universiti Kebangsaan Malaysia station have been used to evaluate the method based on collected datasets for three years, 2011, 2012, and 2013. The experiment findings show that the model has successfully tracked all the up and down patterns in the time series. The results also reveal that VMD-based training might not always provide good results due to the residual signal. Finally, the evaluation focused on generating loss value and comparing it to the ARIMA benchmark. It showed that OVMD-RNN had accomplished a maximum improvement percentage of ARIMA with a value of (99%).

Author 1: Maladh Mahmood Shakir

Author 2: Zalinda Othman

Author 3: Azuraliza Abu Bakar

Keywords: Elman neural networks; forecast; hybrid model; optimized Variational Mode Decomposition; total electronic content

Paper 60: Assaying the Statistics of Crime Against Women in India using Provenance and Machine Learning Models

Abstract: Now-a-days, the surging of crime against women is occurring at a startling rate in India. According to the National Commission for Women, there was a 46% increase in reports of crimes against women in the initial months of the year 2021 in comparison with the same period in 2020. However, to handle this problem, the need of the hour is to fetch relevant and timely information about the various types of crime taking place and make specific predictions based on the existing information to safeguard women from future predictable contingencies. AI and Machine learning mechanisms have become a powerful tool in predicting the crime rate in India under various crime categories by analyzing the crime patterns, crime–centric areas, and the comparative study of various crime categories. Hence, from 2001 to 2019, a women's crime-based dataset from NCRB has been used in this paper, which included various crime sub-categories, for instance; molestation, sexual harassment, rape, kidnapping, dowry deaths, cruelty to family, importation of girls, immortal traffic, sati prevention act, and others. To acquire a better understanding of the data, a framework has been created which makes use of provenance and machine learning algorithms on the dataset, which has been grouped based on several factors such as distribution of cases convicted or reported every year, safest and un-safest states for women in India, etc. Different machine learning algorithms, such as gradient boosting and its many versions, Random forest, and many more, have been used on the dataset. Their performances are evaluated using various metrics such as accuracy, recall, precision, F1 score, and root mean error square.

Author 1: Geetika Bhardwaj

Author 2: R. K. Bawa

Keywords: Crime against women; provenance; scalar techniques; machine learning techniques; decision tree; random forest; gradient boosting; XgBoost; CatBoost; LightGBM

Paper 61: XBLQPS: An Extended Bengali Language Query Processing System for e-Healthcare Domain

Abstract: The digital India program encourages Indian citizens to become conversant with e-services which are primarily English language-based services. However, the vast majority of the Indian population is comfortable with vernacular languages like Bengali, Assamese, Hindi, etc. The rural villagers are not able to interact with the Relational Database Management system in their native language. Therefore, create a system that produces SQL queries from natural language queries in Bengali, containing ambiguous words. This paper proposes a Bengali Query Processor named Extended Bengali language Query Processing System (XBLQPS) to handle queries containing ambiguous words posted to a Healthcare Information database in the electronic domain. The Healthcare Information database contains doctor, hospital and department details in the Bengali language. The proposed system provides support for the Bengali-speaking Indian rural population to efficiently fetch required information from the database. The proposed system extracts the Bengali root word by removing the inflectional part and categorizing them to a specific part of speech (POS) using modified Bengali WordNet. The proposed system uses manually annotated parts of speech detection of a word based on Bengali WordNet. Patterns of noun phrases are generated to detect the correct noun phrase as well as entity and attribute(s). Entity and attributes are used to prepare the semantic table which is utilized to create the Structured Query Language (SQL). The simplified LESK method is utilized to resolve ambiguous Bengali phrases in this query processing system. The accuracy, precision, recall and F1 score of the system is measured as 70%, 74%, 73%, and 73% respectively.

Author 1: Kailash Pati Mandal

Author 2: Prasenjit Mukherjee

Author 3: Atanu Chattopadhyay

Author 4: Baisakhi Chakraborty

Keywords: Relational database management system (RDBMS); modified bengali WordNet; LESK algorithm; structured query language (SQL); natural language query

Paper 62: MSDAR: Multi-Stage Dynamic Architecture Intrusion Detection System

Abstract: Ad hoc networks have been through extensive research in the last decade. Even with their desirable characteristics, major issues related to their security need to be considered. Various security solutions have been proposed to reduce the risks of malicious actions. They mainly focus on key management, authentication, secure localization, and aggregation techniques. These techniques have been proposed to secure wireless communications but they can only deal with external threats. Therefore, they are considered the first line of defense. Intrusion detection systems are always required to safeguard ad hoc networks as such threats cannot be completely avoided. In this paper, we present a comprehensive survey on intrusion detection systems in ad hoc networks. The intrusion detection systems and components and taxonomy as well as different implementations and types of IDSs are studied and categorized. In addition, we provide a comparison between different Intrusion Detection Systems’ architectures. We also propose a Multi Stage Dynamic Architecture intrusion detection system (MSDAR), designed with a multi-stage detection approach making use of both signature-based and anomaly detection benefits. Our proposed intrusion detection system MSDAR is featured by its dynamic architecture as it can be deployed in the network using the Distributed Hierarchical Architecture. The viability and performance of the proposed system MSDAR are tested against the Distributed Denial of Service Attacks through simulations. Advanced performance parameters were used to evaluate the proposed scheme MSDAR. Experimental results have shown that the performance of MSDAR improves by using multiple stages of different detection mechanisms. In addition, based on simulations, the Detection Rate increases when the sensitivity level increases.

Author 1: Ahmed M. ElShafee

Author 2: Marianne A. Azer

Keywords: Ad hoc networks; attacks; DDoS; intrusion detection; security

Paper 63: Data Collection Method for Energy Storage Device of Distributed Integrated Energy Station based on Double Decision Tree

Abstract: The distributed integrated energy station includes an electric energy storage device, heat storage device, cold storage device and other devices. Aiming at the problem of low data acquisition accuracy of energy storage device caused by using a single sensor or acquisition scheme in the existing methods, a new data acquisition method of energy storage device of distributed integrated energy station is designed based on double decision tree algorithm. The data acquisition process of double decision tree algorithm is constructed. On the basis of the process, the mathematical models of electric energy storage device, heat storage device, cold storage device and hybrid energy storage device are established. Then the double decision tree algorithm is used to solve the constructed model, and the acquisition pseudo code is given. So far, the data acquisition of distributed integrated energy station energy storage device based on double decision tree has been completed. The results of case analysis show that the accuracy of this method is higher than 98%, and the collection time is less than 30 ms.

Author 1: Hao Chen

Author 2: Guilian Wu

Author 3: Linyao Zhang

Author 4: Jieyun Zheng

Keywords: Double decision tree; distributed; integrated; energy station; energy storage device; data collection

Paper 64: Reduced False Alarm for Forest Fires Detection and Monitoring using Fuzzy Logic Algorithm

Abstract: The purpose of this research was to detect forest fires in efforts to minimize the incidence of false alarms. The research method used is divided into several stages, namely planning, analysis, design, implementation, testing, and maintenance. Arduino acts as a data collector on the field, which will later be used to detect forest fires and false alarms. Fuzzy logic is used as the essence of the algorithm and will provide a higher level of accuracy for forest fire detection and false alarms. In testing the fuzzy program, the fuzzy output between Arduino and the fuzzy on the monitoring dashboard has a small difference of 0.99%. It can be concluded; the application can minimize the occurrence of false alarms to reduce the user's workload.

Author 1: Maria Susan Anggreainy

Author 2: Bimo Kurniawan

Author 3: Felix Indra Kurniadi

Keywords: False alarm; fuzzy logic; forest fires detection; sensor; microcontroller

Paper 65: Implementation of Gamification in Mathematics m-Learning Application to Creating Student Engagement

Abstract: Mathematics is one of the main subjects in school. In some schools, the learning methods used are still using conventional methods, namely lectures and exercises. The main difficulty in learning mathematics is how to make the material presented more interesting so that it does not make students bored and easy to understand the material. The use of an attitude of interest in games that knows no age and the various advantages of games gives rise to a combination of learning mechanisms called gamification. Gamification is the process of applying game mechanics to non-game activities to increase user interactivity. Gamification in the m-learning mathematics application was developed using the Attention, Relevance, Confidence, and Satisfaction (ARCS) learning model and the octalysis framework gamification method. Gamification in this mathematics m-learning application applies a game strategy using a system of levels, missions, challenges, points, progress bars, leader boards, and badges. The results of this study indicate that this application can be used as an alternative medium for learning mathematics and student engagement with the result that gamification applied to the m-learning mathematics application can increase student interest by 35%, increase student motivation by 33%, and improve understanding 42% of students towards learning mathematics.

Author 1: Sufa Atin

Author 2: Raihan Abdan Syakuran

Author 3: Irawan Afrianto

Keywords: Gamification; m-learning; mathematics; attention; relevance; confidence; and satisfaction (ARCS) model; octalysis framework; student engagement

Paper 66: Energy-based Collaborative Filtering Recommendation

Abstract: The core value of the recommendation model is the using of the measures to measure the difference between the jumps (e.g. pearson), some other studies based on the magnitude of the angle in space (e.g. cosine), or some other studies study the level of confusion (e.g. entropy) between users and users, between items and items. Recommendation model provides an important feature of suggesting the suitable items to user in common operations. However, the classical recommendation models are only concerned with linear problems, currently there is no research about nonlinear problems on the basis of potential/energy approach to apply for the recommendation model. In this work, we mainly focus on applying the energy distance measure according to the potential difference with the recommendation model to create a separate path for the recommendation problem. The theoretical properties of the energy distance and the incompatibility matrix are presented in this article. Two experiment scenarios are conducted on Jester5k, and Movielens datasets. The experiment result shows the feasibility of the energy distance measures/ the potential in the recommendation systems.

Author 1: Tu Cam Thi Tran

Author 2: Lan Phuong Phan

Author 3: Hiep Xuan Huynh

Keywords: Energy distance; energy model; collaborative filtering; recommendation system; distance correlation; incompatibility

Paper 67: Grape Leaves Diseases Classification using Ensemble Learning and Transfer Learning

Abstract: Agriculture remains an important sector of the economy. Plant diseases and pests have a big impact on plant yield and quality. So, prevention and early detection of crop disease are some of the measures that must be implemented in farming to save the plants at an early stage and thereby reduce the overall food loss. Grapes are the most profitable fruit, but they are also vulnerable to a variety of diseases. Black Measles, Black Rot, and Leaf Blight are diseases that affect grape plants. Manual disease diagnosis can result in improper identification and use of pesticides, and it takes a long time. A variety of deep learning approaches have been used to address this issue of the identification and classification of grape leaf diseases. However, there are also limits to such approaches. Therefore, this paper uses deep learning with the concept of ensemble learning based on three famous Convolutional Neural Network (CNN) architectures (Visual Geometry Group (VGG16), VGG19, and Extreme Inception (Xception)). These three models are pre-trained with ImageNet. The performance of the proposed approach is analyzed using the Plant Village (PV) dataset of common grape leaf diseases. The Proposed model gives higher performance than the results achieved by using each Deep Learning architecture separately and compared with the recent approaches in this study. The proposed system outperformed the others with 99.82% accuracy.

Author 1: Andrew Nader

Author 2: Mohamed H.Khafagy

Author 3: Shereen A. Hussien

Keywords: Ensemble learning; grape leaf diseases; convolutional neural network (CNN); transfer learning

Paper 68: An Extractive Text Summarization based on Candidate Summary Sentences using Fuzzy-Decision Tree

Abstract: This study aims to predict candidate summary sentences in extractive summary using the Fuzzy-Decision Tree method. The fuzzy method is quite superior and the most widely used in extractive summaries, because Fuzzy has advantages in calculations that are not cryptic, so it is able to calculate uncertain possibilities. However, in its implementation, the fuzzy rule generation process is often carried out randomly or based on expert understanding so that it does not represent the distribution of the data. Therefore, in this study, a Decision Tree (DT) technique was added to generate fuzzy rules. From the fuzzy final result, important sentences are obtained that are candidates for summary sentences. The performance of our proposed method was tested on the 2002 DUC dataset in the ROUGE-1 evaluation. The results showed that our method outperformed other methods (baseline and sentence ranking) with an average precision of 0.882498, Recall 0.820443 and F Measure 0.882498 with CI for F1 0.821-0.879 at the 95% confidence level.

Author 1: Adhika Pramita Widyassari

Author 2: Edy Noersasongko

Author 3: Abdul Syukur

Author 4: Affandy

Keywords: Text summarization; extractive; fuzzy; decision tree

Paper 69: Design and Implementation of ML Model for Early Diagnosis of Parkinson’s Disease using Gait Data Analysis in IoT Environment

Abstract: Parkinson’s disease (PD) is the world’s second most neurodegenerative disorder that results in a steady loss of movement. The symptoms in patients occur slowly with the passage of time are and very hard to identify in its initial stage. So, early diagnosis of PD is the foremost need for timely treatment to people. The introduction of smart technologies like the Internet of Things (IoT) and wearable sensors in the healthcare domain offers a smart way of identifying the symptoms of PD patients. In which smart sensors are worn on the patient’s body which continuously monitor the symptoms in patients and track their possible health status. The major objective of this work is to propose a machine learning-based healthcare model that best classifies the subjects into healthy and Parkinson's patients by extracting the most important features. A step regression-based feature selection method is followed to improve the classification of PD. A Shapiro Wilk test is adopted to check the normality of the gait dataset. This model is implemented on three publicly available Parkinson’s datasets collected from three different studies available on Psyionet. All these data sets contain VGRF recordings obtained from eight different sensors placed under each foot. Experimentation is done on the Jupyter notebook by utilizing Python as a programming language. Experimental results revealed that our proposed model with effective pre-processing, feature extraction, and feature selection method achieved the highest accuracy result of 95.54%, 98.80%, and 94.52% respectively when applied to three datasets. Our research inducts knowledge about significant characteristics of a patient suffering from PD and may help to diagnose and cure at an early stage.

Author 1: Navita Mehra

Author 2: Pooja Mittal

Keywords: Internet of things (IoT); sensors; parkinson's disease (PD); machine learning (ML); vertical ground reaction force (VGRF)

Paper 70: A Novel Methodology for Disease Identification Using Metaheuristic Algorithm and Aura Image

Abstract: Every human has a specific Aura. Every organism in the human body emits energy comprising of Ultra Violet radiation, thermal radiation, and electromagnetic radiation. These energy levels help to underline the physical health inside the human body. In general, these energy levels are called Aura. In order to capture the energy levels, specific cameras like Kirlian are used. These cameras try to capture the energy distribution and map them to the individual organs of the human body. In this article, we present a methodology using Image processing techniques, where Bivariate Gaussian Mixture Model (BGMM) is considered as a classifier to identify the diseases in humans based on the energy distribution. In this article, we have considered five different categories of diseased organs that are identified based on the energy distribution. The preprocessing is subjected to the morphological technique and Particle Swarm Optimization (PSO) algorithm is considered for feature extraction. The segmentation process is carried out using the feature extracted and training is carried out using the BGMM classifier. The result obtained is summarized using various other methods like Support Vector Machine (SVM), Artificial Neural Network (ANN), and Multiclass SVM (MSVM). The results showcase that the proposed methodology exhibits recognition accuracy at 90%.

Author 1: Manjula Poojary

Author 2: Yarramalle Srinivas

Keywords: Aura images; BGMM; image classification; multiclass SVM; artificial neural network

Paper 71: Enhanced Gradient Boosting Machines Fusion based on the Pattern of Majority Voting for Automatic Epilepsy Detection

Abstract: Automatic detection of epilepsy based on EEG signals is one of the interesting fields to be developed in medicine to provide an alternative method for detecting epilepsy. High accuracy values are very important for accurate diagnosis in detecting epilepsy and avoid errors in diagnosing patients. Therefore, this study proposes the Enhanced Gradient Boosting Machines Fusion (Enhanced GBM Fusion) for automatically detecting epilepsy based on electroencephalographic (EEG) signals. Enhanced part of GBM Fusion is the pattern of majority voting evaluation based on the fusion of five-class and two-class GBM, called Enhanced GBM Fusion. The raw signal is extracted using Discrete Fourier Transform (DFT) and Discrete Wavelet Transform (DWT), then feature is selected by using Genetic Algorithm (GA) before classification. This proposed method was evaluated using five classes (normal in open eyes, normal in close eyes, interictal with hippocampal, interictal, and ictal) from the University of Bonn. The experimental results show that the proposed Enhanced GBM Fusion can increase the accuracy of GBM Fusion of 99.8% to classify five classes of epilepsy based on EEG signal. However, the performance of Enhanced GBM Fusion cannot be generalized to other datasets.

Author 1: Dwi Sunaryono

Author 2: Riyanarto Sarno

Author 3: Joko Siswantoro

Author 4: Diana Purwitasari

Author 5: Shoffi Izza Sabilla

Author 6: Rahadian Indarto Susilo

Author 7: Adam Abelard Garibaldi

Keywords: Epilepsy; enhanced gradient boosting machine fusion; electroencephalographic (EEG) signal; discrete wavelet transform (DWT); discrete fourier tansform (DFT); genetic algorithm (GA)

Paper 72: Intelligent Framework for Enhancing the Quality of Online Exams based on Students’ Personalization

Abstract: In education sector, personalization is an evolutionary term that gained a high attention due to its effectiveness in raising the enterprise competence level. This research aims at proposing a novel model for effective smart testing, which considers the student’s Facebook activities in determining the students’ personality and constructing his suitable exam. The aim of this examination perspectives to ensure the reliable student evaluation according to his gained knowledge to ensure that no other factor interferes which may negatively affect the reliable evaluation. The research also applies text analytics techniques to ensure the exam balance. The proposed model has been applied and evaluated with professors’ percentage equal to 96.5 % and successfully reach students satisfaction percentage with average equal to 96.63%.

Author 1: Ayman E. khedr

Author 2: Abdulwahab Ali Almazroi

Author 3: Amira M. Idrees

Keywords: Personalization; data mining; sentiment analysis; social networks; e-learning

Paper 73: Churn Prediction Analysis by Combining Machine Learning Algorithms and Best Features Exploration

Abstract: The market competition and the high cost of acquiring new customers have led financial organizations to focus more and more on effective customer retention strategies. Although the banking and financial sectors have low churn rates compared to other sectors, the impact on profitability related to losing a customer is comparatively high. Thereby, customer turnover management and analysis play an essential part for financial organizations in order to improve their long-term profitability. Recently, it appears that using machine learning to predict churning improves customer retention strategies. In this work, we discuss some specific machine learning models proposed in the literature that deal with this problem and compare them with some emerging models, based on Ensemble learning algorithms. As a result, we build a predictive churn approaches that look at the customer history data, check to see who is active after a certain time and then create models that identify stages where a customer can leave the concerned company service. Ensemble learning algorithms are also used to find relevant features in order to reduce their number which is of great importance when performing the training step with some classical models such us Multi-Layer Perception Neural networks. The proposed approaches can achieve up to 89% in accuracy when other research works, dealing with the same dataset, can achieve less than 86%.

Author 1: Yasyn ELYUSUFI

Author 2: M’hamed AIT KBIR

Keywords: Customer churn; prediction; machine learning

Paper 74: Rider Driven African Vulture Optimization with Multi Kernel Structured Text Convolutional Neural Network for Classifying e-Commerce Reviews

Abstract: Opinion mining is a natural language processing based on sentiment classification technique to determine the sentiment of the reviews. The major existing text Convolutional Neural Network (CNN) algorithms are derived based on 3×3 size kernels which extract ineffective review text-features and lead to less classification accuracy. Moreover, most of the traditional CNN versions output three classes such as positive, negative, and neutral as their classification results. Hence, a novel algorithm namely ‘RAVO driven Multi-Size Kernel structured Text CNN for classifying ecommerce reviews (MSK-TCNN-RAVO)’ is proposed in this work. This proposed approach utilizes five multi-size kernels (3×7,5×7,1×3,1×5,1×7), multi-dimensional kernels (1D & 2D), and integrates varying size kernels to extract text-features effectively. In addition, the performance of multi-kernel CNN is highly enhanced by RAVO algorithm based on rider optimization. Moreover, the proposed approach is highly effective to process 'review-stop-words removal' that decrease the complexity and time consumption of the opinion mining process. Most existing systems use single pooling operations which reduce feature map processing performance, hence, dual pooling operations (both Max and Average pooling) are employed in this research. Furthermore, it is configured to generate five classification outputs such as bad, fair, neutral, good, and excellent to support better decision-making with 95.5% accuracy. This method is evaluated using different quality metrics and five review-databases to measure the performance, and the results reveal that the proposed method outperforms the other existing review classification algorithms.

Author 1: H. Mohamed Zakir

Author 2: S. Vinila Jinny

Keywords: Natural language processing; opinion mining; convolutional neural network; text sentiment classification; ecommerce review

Paper 75: Comparison of Image Enhancement Algorithms for Improving the Visual Quality in Computer Vision Application

Abstract: Computer vision has its numerous real-world applications on Visual Object Tracking which includes human-computer interaction, autonomous vehicles, robotics, motion-based recognition, video indexing, surveillance and security, human-computer interaction, autonomous vehicles, robotics, motion-based recognition, video indexing, surveillance and security. The factors affecting the tracking process is due to low illumination, haze and cloudy environment and noisy environment. In this paper, we aim to extensively review the latest trends and advances in adaptive enhancement algorithm and evaluate the performance using Full reference like, SSIM (Structure Similarity Index Measure), MS-SSIM (Multi-scale Structure Similarity Index Measure), ESSIM (Edge Strength Structural Similarity Index), FSIM (Feature Similarity Index Measure), VIF (Visual Information Fidelity), CW-SSIM (complex wavelet structural similarity), UQI (Universal Quality Index), IEF (Image Enhancement Factor), IQI (Image Quality Index), EME (Enhancement Measurement Error), CVSI (Contrast and Visual Salient Information), MCSD (Multiscale contrast similarity deviation), NQM (Noise Quality Measure), Gradient Magnitude Similarity Mean (GMSM), Gradient Magnitude Similarity Deviation (GMSM) and no-reference image quality measures Perception based Image Quality Evaluator (PIQE), Blind/Reference less Image Spatial Quality Evaluator (BRISQUE), Naturalness Image Quality Evaluator (NIQE), Average Gradient (AG), Contrast, Information Entropy (IE), Lightness order Error (LOE). The main purpose of adaptive image enhancement is to smooth the uniform area and sharpen the border of an image to improve its visual quality. In this paper, fourteen image enhancement algorithms were tested on LoL dataset to benchmark the time taken to process them and their output quality was evaluated. Results from this study will give insights to image analysts for selecting image enhancement algorithms which acts as a pre- processing stage for Visual object Tracking.

Author 1: Jenita Subash

Author 2: Jharna Majumdar

Keywords: Tracking; robotics; surveillance; enhancement

Paper 76: Effect of Feature Engineering Technique for Determining Vegetation Density

Abstract: Vegetation density is one type of information collected from vegetation cover. Vegetation density influences evapotranspiration in terrain, which is essential in assessing how vulnerable peatlands are to fire. The Keetch and Byram Drought Index model, which evaluates peatland fire vulnerability, divides vegetation density into heavily grazed, softly grazed, and un-grazed. Manual approaches for analyzing vegetation density in the field, on the other hand, need a significant amount of resources. Image data acquisition, pre-processing, feature extraction, classification, feature selection, classification, and validation are all computer vision approaches used to solve these problems. Artificial intelligence algorithms and machine learning approaches promise outstanding accuracy in modern computer vision research. However, in the classification process, the impact of feature extraction is critical. Pattern identification at Back Propagation Neural Network (BPNN) is problematic because the feature extraction dimension is excessively complicated. The solution to this problem is using the feature engineering technique to choose the characteristics. This research aims to explore how feature engineering influences the accuracy of results. According to the statistics, implementing the recommended strategy can increase accuracy by 1% and increase kappa by 1.5%. This increase in vegetation density classification accuracy might help detect peatland vulnerability sooner. The novel aspect of this paper is that, after feature extraction, a feature engineering strategy is used in the machine learning classification stage to reduce the number of complex dimensions.

Author 1: Yuslena Sari

Author 2: Yudi Firmanul Arifin

Author 3: Novitasari Novitasari

Author 4: Mohammad Reza Faisal

Keywords: Vegetation cover; vegetation density; feature extraction; feature engineering; accuracy

Paper 77: Techno-pedagogical Solution to Support the Improvement of the Quality of Education in Technical and Vocational Training in Mauritania

Abstract: E-learning is the most promising and fastest growing activity since the advent of the COVID-19 pandemic. Although the pandemic seems to have been eradicated in several countries of the world, it is worth mentioning that some positive cases of the covid-19 variant have been detected. This could accelerate the rate of infection again. Hence the interest generate in reinforcing the quality of distance learning platforms. Technical and vocational training (TVT) in Mauritania is based on science, technology, engineering, and mathematics (STEM) disciplines. Unfortunately, the expansion of the COVID-19 pandemic has negatively impacted the quality of education with a halt in teaching affecting 8000 students. Yet, the quality of education in these disciplines is a key factor in meeting the demands of emergence and economic growth. This paper advocates the mixed pedagogical model by proposing a techno-pedagogical solution to improve the quality of teaching and learning processes. The proposed solution combines the use of technologies such as Modular Object-Oriented Dynamic Learning Environment (Moodle) and Web Real Time Communication (WebRTC) to provide pedagogical services in a context with a limited Internet connection. In addition, we set up a signaling system to maintain direct communication between the pairs, Application Programming Interface (API) of Multipoint Control Unit (MCU) to ensure simultaneous collaboration in a peer-to-peer context, used implementations of security protocols such as Datagram Transport Layer Security (DTLS) and Secure Real-time Transport Protocol (SRTP) to secure data transport.

Author 1: Cheikhane SEYED

Author 2: Jeanne Roux NGO BILONG

Author 3: Mohamed Ahmed SIDI

Author 4: Mohamedade Farouk NANNE

Keywords: TVT; STEM; Mixed education; Moodle; WebRTC; signaling system; MCU; DTLS; SRTP

Paper 78: Students’ Characteristics of Student Model in Intelligent Programming Tutor for Learning Programming: A Systematic Literature Review

Abstract: This study describes preliminary results of a research related to Intelligent Programming Tutor (IPT) which is derived from Intelligent Tutoring System (ITS). The system architecture consists of four models. However, in this study student model mainly student characteristic was focused. From literature, 44 research articles were identified from a number of digital databases published between 1997 to 2022 base on systematic literature review (SLR) method. The findings show that the majority 48% of IPT implementation focuses on knowledge and skills. While 52% articles focused on a combination of two to three student characteristics where one of the combinations is knowledge and skill. When narrow down, 25% focused on knowledge and skills with errors or misconceptions; 4% focused on knowledge and skill with cognitive features; 5% focused focus on knowledge and skill with affective features; 2% focused on knowledge and skill with motivation; and 9% based on knowledge and skill with learning style and learning preferences as students’ characteristics to build their student model. Whereas 5% focused on a combination of three student characters which are knowledge and skill with cognitive and affective features and 2% focused on knowledge and skill with learning styles and learning preferences and motivation as students’ characteristics to construct the tutoring system student model. To provide an appropriate tutoring system for the students, students’ characteristic needs to decide for the student model before developing the tutoring system. From the findings, it can say that knowledge and skills is an essential students’ characteristic used to construct the tutoring system student model. Unfortunately, other students’ characteristic is less considered especially students’ motivation.

Author 1: Rajermani Thinakaran

Author 2: Suriayati Chuprat

Keywords: Intelligent tutoring system; intelligent programming tutor; student characteristics; student model

Paper 79: News Analytics for Business Sentiment Suggestion

Abstract: Business and economics news has become one of the factors businesses consider when making decisions. However, the exponential increase in the availability of business information sources on the internet makes it more difficult for entrepreneurs to keep up with and extract useful insights from many news articles. Although many preceding works focused on the sentiment extracted in the news, the results were intended for everyone. The sentiments based on a user's queries are needed to provide customized service. Hence, this paper proposed a system integrated into a chatbot to automatically understand users' queries and recommend sentiments based on news articles. The main objective is to provide entrepreneurs, especially those considering international trade and investment, with the sentiments embodied in the latest news articles to help them keep up with the business and economic trends relevant to them. The methodology is based on deep learning and transfer learning. A pre-trained deep learning model was fine-tuned for natural language processing tasks to perform sentiment analysis in news articles. A survey questionnaire was used to measure the effectiveness of the system. The survey result showed that most users agreed with the predicted sentiments from the system.

Author 1: Sirinda Palahan

Keywords: Sentiment analysis; deep learning; pre-trained model; natural language processing

Paper 80: Sentiment Analysis using Term based Method for Customers’ Reviews in Amazon Product

Abstract: Customers’ review in Amazon platform plays an important role for making online purchase decision making, however the reviews are snowballing in E-commerce day by day. The active sharing of customers’ experience and feedback helps to predict the products and retailers’ quality by using natural language processing. This paper will focus on experimental discussion on Amazon products reviews analysis coupled with sentiment analysis using term-based method and N-gram to achieve best findings. The investigation of sentiment analysis on amazon product gain more valuable information on related text to solve problem related services, products information and quality. The analysis begins with data pre-processing of Amazon products reviews then feature extraction with POS tagging and term-based concept. e-Commerce customer’s reviews normally classify different experience into positive, negative and neutral to judge human behavior and emotion towards the purchase products. The major findings discussed in this journal will be using four different classifier and N-grams methods by computing accuracy, precision, recall and F1-Score. TF-IDF method with N-gram shows unigram with Support Vector Machine learning with highest accuracy results for Amazon product customers’ reviews. The score reveals that Support Vector Machine for unigram achieved 82.27% for accuracy, 82% precision, 80% Re-call and 72% F1-Score.

Author 1: Thilageswari a/p Sinnasamy

Author 2: Nilam Nur Amir Sjaif

Keywords: Sentiment analysis; e-commerce; term based; n-gram

Paper 81: An Iootfuzzer Method for Vulnerability Mining of Internet of Things Devices based on Binary Source Code and Feedback Fuzzing

Abstract: With the technological progress of the Internet and 5G communication network, more and more Internet of Things devices are used in it. Limited by the cost, power consumption and other factors of Internet of Things devices, the systems carried by the Internet of Things devices often lack the security protection provided by larger equipment systems such as desktop computers. Because the current personal computers and servers mostly use the x86 architecture, and the previous research on security tools or hardware-based security analysis feature support is mostly based on the x86 architecture, the traditional security analysis techniques cannot be applied to the current large-scale ARM-based and MIPS-based Internet of Things devices. Based on this, this paper studies the firmware binary program of common Linux-based Internet of Things devices. A binary static instrumentation technology based on taint information analysis is proposed. The paper also analyzes how to use the binary static instrumentation technology combined with static analysis results to rewrite binary programs and obtain taint path information when binary programs are executed. Firmware binary fuzzing technology based on model constraints and path feedback is studied to cover more dangerous execution paths in the target program. Finally, iootfuzzer, a prototype vulnerability mining system for firmware binaries of Internet of Things devices, is used to test and analyze the two technologies. The results show that its fuzzing efficiency for Internet of Things devices is better than other fuzzing technologies such as boofuzz and Peach 3. It can fill in some gaps in the current security analysis tools for the Internet of Things devices and improve the efficiency of security analysis for Internet of Things devices, which contributes to the field through automated security vulnerability detection systems.

Author 1: Guangxin Guo

Author 2: Chao Wang

Author 3: Jiahan Dong

Author 4: Bowen Li

Author 5: Xiaohu Wang

Keywords: Internet of things; system vulnerabilities; source code; fuzz testing; instrumentation technology

Paper 82: Mobile Learning in Science Education to Improve Higher-Order Thinking Skills (HOTS) and Communication Skills: A Systematic Review

Abstract: Today, the increasing use of technology and mobile applications in education was interesting. This research was a systematic review study with limited 30 articles from 2012 to 2021. It aims to answer research questions about what mobile devices used in learning and what learning approaches are used in science learning to improve higher-order thinking skill and communication skills. The findings of this study were in line with the research objectives: First, the most appropriate mobile devices used to achieve learning objectives are mobile phones, followed by PDAs, tablets, iPad, laptops, e-books, and iPods. Second, the learning approach used in science learning to improve higher-order thinking skills and communication skills are a collaborative learning approach, inquiry learning, project-based learning, problem-based learning, game-based learning, and flipped classroom learning. It was hoped that this research can be an illustration for other researchers to create innovative learning approaches. Some research that can be done next based on this research is how mobile learning in social learning or comparing the two, further research on the most appropriate learning media for mobile learning, and research on the effectiveness of implementing approach strategies in mobile learning.

Author 1: Adilah Afikah

Author 2: Sri Rejeki Dwi Astuti

Author 3: Suyanta Suyanta

Author 4: Jumadi Jumadi

Author 5: Eli Rohaeti

Keywords: Communication skills; higher-order thinking skills; mobile learning; science education

Paper 83: Superpixel Sizes using Topology Preserved Regular Superpixel Algorithm and their Impacts on Interactive Segmentation

Abstract: Interactive Image Segmentation is a type of semi-automated segmentation that uses user input to extract the object of interest. It is possible to speed up and improve the end result of segmentation by using pre-processing steps. The use of superpixels is an example of a pre-processing step. A superpixel is a collection of pixels with similar properties such as texture and colour. Previous research was conducted to assess the impact of the number of superpixels (based on SEEDS superpixel aglorithms) required to achieve the best segmentation results. The study, however, only examined one type of input (strokes) and a small number of images. As a result, the goal of this study is to extend previous work by performing interactive segmentation with input strokes and a combination of bounding box and strokes on images from Grabcut image data sets generated by Topology preserved regular superpixel (TPRS). Based on our findings, an image with 1000 to 2500 superpixels and a combination of bounding box and strokes will help the interactive segmentation algorithm produce a good segmentation result. Finally, the size of the superpixels would influence the final segmentation results as well as the input type.

Author 1: Kok Luong Goh

Author 2: Soo See Chai

Author 3: Giap Weng Ng

Author 4: Muzaffar Hamzah

Keywords: Image segmentation; superpixel; input type; interactive segmentation

Paper 84: Design of Intelligent Fusion Terminal System with Fog Computing Capability in Distribution Area based on Large Capacity CPU

Abstract: The intelligent fusion terminal in the distribution area usually adopts the mode of cooperation between the cloud and the edge, and the workload of manual operation and maintenance is large. Therefore, an intelligent fusion terminal system in the distribution area with fog computing capability based on a high-capacity CPU is proposed. Follow the “cloud pipe edge end” construction framework of smart IOT system, and take this framework as the edge computing node of distribution station area and power consumption side. Mt7622b chip in Linux operating system with openwrt firmware is used as the main control chip of edge agent gateway equipment, and the recursive least square method is used to realize the data fusion of power acquisition service in distribution station area and power distribution demand terminal. The test results show that the designed system can realize real-time monitoring of power consumption and distribution data and power quality management in the distribution station area, and the data processing delay is less than 100ms, which provides a reference for the intelligent fusion terminal system in the distribution station area.

Author 1: Ou Zhang

Author 2: Songnan Liu

Author 3: Hetian Ji

Author 4: Xuefeng Wu

Author 5: Xue Jiang

Keywords: Large-capacity CPU; fog computing capability; power distribution station area; intelligent integration; terminal; system design

Paper 85: A Blind Robust Image Watermarking on Selected DCT Coefficients for Copyright Protection

Abstract: This paper proposes a blind and robust image watermarking technique using Discrete Cosine Transform (DCT) for copyright protection on color images called BRIW-DCT. Each channel of the host image is divided into non-overlapping image blocks with the size of 8×8 pixels. Each image block is transformed into a frequency domain using the DCT transformation. The watermark image is embedded into the host image by modifying the 11th to the 15th DCT coefficient. The experimental result shows that the watermarked image achieved a high PSNR value of 50.4489 dB and a high SSIM value of 0.9991. Furthermore, various attacks are performed on the watermarked image. BRIW-DCT can successfully recover the watermark image from the tampered image, which produces a high NC value of 0.7805 and a low BER value of 0.1126.

Author 1: Majid Rahardi

Author 2: Ferian Fauzi Abdulloh

Author 3: Wahyu Sukestyastama Putra

Keywords: Robust watermarking; copyright protection; discrete cosine transform; frequency domain; color image watermarking

Paper 86: Feedback Model when Applying the Evaluation by Indicators in the Development of Competences through Problem based Learning in a Systems Engineering Course

Abstract: Feedback can be very influential in students' learning, therefore, the university must be very clear about its procedures and rules on the time lapse of the response of the work done by them, and the comments made to positively influence in a sustained manner in the evaluation of learning. The present work shows the experience of applying the Problem Based Learning (PBL) methodology and also developing research competencies through Formative Research, and as a result of the evaluation of the learning of the Criteria and its Performance Indicators corresponding to the course Business Electronic which is taught by two teachers in theory and laboratory practices. The objective is to design a Feedback Model for the problems solved by the students in order to support the improvement of their learning. The methodology used is Problem Based Learning together with the Feedback Model, of real problems posed contemplating different contexts of the organisations; we have that from the Deliverable Report of each problem at the same time the incidences and observations are registered in the corresponding register and in this way the Feedback Report is elaborated. The results obtained reveal that the objectives of producing the Feedback Report are achieved, which should be sent as soon as possible to the students for analysis, to propose their own strategies for improving the shortcomings or errors, as well as having the motivation to continue progressing by accepting the suggestions or contributions of the teacher; as well as seeing an increase in knowledge, development of their competences, skills, attitudes, making their own judgements, and achieving the Student Results. In conclusion, the application of a well-planned active didactic strategy, the adequate evaluation of learning through the qualification of the indicators of each criterion, and the elaboration of a timely feedback report on the problems, will achieve the expected results for both the course and the student.

Author 1: César Baluarte-Araya

Author 2: Oscar Ramirez-Valdez

Keywords: Problem based learning; competencies; evaluation; criteria; performance indicator; deliverable report; feedback report; skills; formative inquiry

Paper 87: Development of Discrepancy Evaluation Model based on Tat Twam Asi with TOPSIS Calculation

Abstract: This research had the main objective to provide information related to the innovation available in the form of an educational evaluation model that integrates the Discrepancy evaluation component, Tat Twam Asi concept, and TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method in the framework of determining the dominant indicators triggering the effectiveness of implementing blended learning in IT vocational schools. The approach of this research was development research by an R & D development model that focused on four stages, including a) research and field data collection, b) planning, c) design development, d) initial trial, and e) revisions to the results of the initial trial. There were 34 subjects involved in the trial design of the evaluation model in this research, including two education experts, two informatics experts, and 30 IT vocational teachers in Bali. The instruments used in data collection were in the form of questionnaires, interview guidelines, and photo documentation. The analysis technique for the data that had been collected used quantitative descriptive techniques that referred to percentage descriptive calculations. The results of this research were Tat Twam Asi-based Discrepancy evaluation model design which was integrated with TOPSIS calculations and had been classified as excellent according to the eleven-scale categorization table.

Author 1: Dewa Gede Hendra Divayana

Author 2: Agus Adiarta

Author 3: P. Wayan Arta Suyasa

Keywords: Discrepancy; evaluation model; tat twam asi; TOPSIS

Paper 88: Solving the Job Shop Scheduling Problem by the Multi-Hybridization of Swarm Intelligence Techniques

Abstract: The industry is subject to strong competition, and customer requirements which are increasingly strong in terms of quality, cost, and deadlines. Consequently, the companies must improve their competitiveness. Scheduling is an essential tool for improving business performance. The production scheduling problem is usually an NP-hard problem, its resolution requires optimization methods dedicated to its degree of difficulty. This paper aims to develop multi-hybridization of swarm intelligence techniques to solve job shop scheduling problems. The performance of recommended techniques is evaluated by applying them to all well-known benchmark instances and comparing their results with the results of other techniques obtainable in the literature. The experiment results are concordant with other studies that have shown that the multi hybridization of swarm intelligence techniques improve the effectiveness of the method and they show how these recommended techniques affect the resolution of the job shop scheduling problem.

Author 1: Jebari Hakim

Author 2: Siham Rekiek

Author 3: Kamal Reklaoui

Keywords: Scheduling; Job shop; Multi-hybridization; Swarm intelligence

Paper 89: An IoT-based Fire Safety Management System for Educational Buildings: A Case Study

Abstract: Safety is a serious concern that should be addressed carefully in different locations including homes, workplaces and educational buildings. The risk of fire is the most significant threat in many educational facilities such as schools, universities, offices, etc. The main goal of this work is developing an effective system that allows early managing of fires to avoid material and human losses. With the advent of the Internet of Things (IoT), the implementation of such system became possible. A low-cost system incorporating IoT sensors is constructed in this study to collect data (heat, the number of people at the fire scene, ...) in real time. The system provides a control panel that displays readings from all sensors on a single web page. When the collected values exceed a particular threshold, the system sends a message to the building keeper’s phone, allowing him to notify the authorities or dispatch firemen in real time. One of the system’s most important characteristics is that it keeps track of how many people are at the fire scene, simplifying the evacuation process and allowing civil defense authorities to efficiently manage resources. The system has been successfully tested in a variety of circumstances in an educational building (Al-Faisaliah female campus, University of Jeddah, Saudi Arabia).

Author 1: Souad Kamel

Author 2: Amani Jamal

Author 3: Kaouther Omri

Author 4: Mashael Khayyat

Keywords: Safety; fire; Internet of Things (IoT); sensors; cloud based platform; ThingSpeak

Paper 90: Structural Vetting of Academic Proposals

Abstract: Increasing postgraduate enrollments gives rise to many proposal documents required for vetting and human supervision. Reading and comprehension of large documents is a boring and somewhat difficult task for humans which can be delegated to machines. One way of assisting supervisors with this routine screening of academic proposals is to provide an artificial intelligent (AI) tool for initial structural vetting— checking if sections of proposals are complete and appear where they are supposed to. Natural Language Processing (NLP) techniques in AI for document vetting has been applied in legal and financial domains. However, in academia, available tools only perform tasks such as checking proposals for plagiarism, spellings or grammar, word editing, and not structural vetting of academic proposal. This paper presents a tool named Auto-proofreader that attempts to perform the task of structural document review of proposals on behalf of the human expert using formal techniques and document structure understanding hinged on context free grammar rules (CFGs). The experimental results on a corpus of 20 academic proposals using confusion matrix technique for evaluation gives an overall of 87% accuracy. This tool is expected to be a useful aid in postgraduate supervision for vetting students’ academic proposals.

Author 1: Opeoluwa Iwashokun

Author 2: Abejide Ade-Ibijola

Keywords: Document structure; context free grammar; post-graduate supervision; artificial intelligence; natural language pro-cessing

Paper 91: Mobile Application Prototype: Learning in the Programming Course in Computer Engineering Students

Abstract: Students need to continue with the learning process related to the world of programming because today are in the era of technological globalization. Therefore, it is very important to learn about it, since programming is used in different areas and as a result obtain software, electronic devices, among others. seek to design a mobile application that helps students learn much more about programming, since students in the first cycles of computer science and computer science have difficulties learning about different programming languages. That is why the application seeks to help the student by complementing their learning in such a way that they can obtain favorable results in their progress thanks to the development of the application. The objective is to design a mobile application for teaching programming in a didactic way that helps computer science students with learning difficulties. The methodology used is Design Thinking, because it is an agile methodology that is based on phases that help us understand and collect information about the problem encountered in order to provide a solution. As for the case study, the design of the mobile application and the detailed development of the prototype are shown. The result obtained is the prototype of the mobile application in which students with learning difficulties will benefit. In addition, a survey carried out at the University of Sciences and Humanities to students and teachers is shown, where very relevant data is obtained according to their learning.

Author 1: Lilian Ocares-Cunyarachi

Author 2: Laberiano Andrade-Arenas

Keywords: Design thinking; learning; mobile application; stu-dents; programming

Paper 92: Prediction of COVID-19 Patients Recovery using Ensemble Machine Learning and Vital Signs Data Collected by Novel Wearable Device

Abstract: During the spread of a pandemic such as COVID- 19, the effort required of health institutions increases dra-matically. Generally, Health systems’ response and efficiency depend on monitoring vital signs such as blood oxygen level, heartbeat, and body temperature. At the same time, remote health monitoring and wearable health technologies have revolutionized the concept of effective healthcare provision from a distance. However, analyzing such a large amount of medical data in time to provide the decision-makers with necessary health procedures is still a challenge. In this research, a wearable device and monitoring system are developed to collect real data from more than 400 COVID-19 patients. Based on this data, three classifiers are implemented using two ensemble classification techniques (Adaptive Boosting and Adaptive Random Forest). The analysis of collected data showed a remarkable relationship between the patient’s age and chronic disease on the one hand and the speed of recovery on the other. The experimental results indicate a highly accurate performance for Adaptive Boosting classifiers, reaching 99%, while the Adaptive Random Forest got a 91% accuracy metric.

Author 1: Hasan K. Naji

Author 2: Hayder K. Fatlawi

Author 3: Ammar J. M. Karkar

Author 4: Nicolae GOGA

Author 5: Attila Kiss

Author 6: Abdullah T. Al-Rawi

Keywords: Machine learning; COVID-19; wearable device

Paper 93: Mobile Application Design: Sale of Clothes Through Electronic Commerce

Abstract: During the COVID-19 pandemic, small clothing sales companies lost economic income and customers due to a lack of digital transformation, causing the dismissal of many employees. Due to this problem, our objective is to design an e-commerce mobile application for the sale of clothes, so that Small and medium-sized enterprises dedicated to this area generate income and retain their customers. For this, the Rational Unified Process (RUP) methodology was applied, because this method-ology provides a structured way for companies or developers to visualize the development of the software, and for the validation by expert judgment, the survey and the questionnaire were used as instruments. Obtaining as a result a positive rating for the design of the mobile application and its acceptance to accommodate what is reflected. In conclusion, the e-commerce mobile application was successfully designed, backed by expert judgment, so that Small and medium-sized enterprises can offer their products and generate income, as well as build customer loyalty

Author 1: Raul Jauregui-Velarde

Author 2: Franco Gonzalo Conde Arias

Author 3: Jose Luis Herrera Salazar

Author 4: Michael Cabanillas-Carbonell

Author 5: Laberiano Andrade-Arenas

Keywords: Mobile application; COVID-19; e-commerce; RUP; sale of clothes

Paper 94: Comparative Analysis of Machine Learning Algorithms and Data Mining Techniques for Predicting the Existence of Heart Disease

Abstract: Heart diseases are considered one of the leading causes of death globally over the world. They are difficult to be predicted by a specialist physician as it is not an easy task which requires greater knowledge and expertise for prediction. With the variety of machine learning and deep learning algorithms, there exist many recent studies in the state of the art that have been done remarkable and practical works for predicting the presence of heart diseases. However, some of these works were affected by various drawbacks. Hence, this work aims to compare and analyze different classifiers, pre-processing, and dimensionality reduction techniques (feature selection and feature extraction) and study their effect on the prediction of heart diseases existence. Therefore, based on the resulting performance of several conducted experiments on the well-known Cleveland heart disease dataset, the findings of this study are: 1) the most significant subset of features to predict the existence of heart diseases are PES, EIA, CPT, MHR, THA, VCA, and OPK, 2) Naïve Bayes classifier gave the best performance prediction, and 3) Chi-squared feature selection was the data mining technique that reduced the number of features while maintained the same improved performance for predicting the presence of heart disease.

Author 1: Nourah Alotaibi

Author 2: Mona Alzahrani

Keywords: Heart disease; feature selection; feature extraction; dimensionality reduction; Chi-squared; Naive Bayes; Cleveland dataset

Paper 95: Crop Field Monitoring and Disease Detection of Plants in Smart Agriculture using Internet of Things

Abstract: The Internet of Things can be defined as the network of physical objects that have sensors, software, and other technologies built into them in order to communicate and exchange data with other systems and devices over the internet. In intelligent agricultural advancements to increase the quality of agriculture, the Internet of Things (IoT) can be used. The manual monitoring of plant diseases is quite challenging. It demands enormous effort, expertise in the diseases of plants and the considerable time required for processing. The idea of automation in Smart Agriculture is implemented using the Internet of Things (IoT). They help monitor the plant leaf conditions, control water irrigation, gather images using installed IoT system which includes NodeMCU, cameras, soil moisture, temperature sensors and detect diseases in plants on the datasets collected from leaves. To detect plant diseases, image processing is applied. The detection of diseases comprises the acquisition of images, image pre-processing, segmenting an image, extracting and classifying features. In addition, the performance of two machine-learning techniques, such as a linear and polynomial kernel multi hidden extreme machine (MELM) and a support vector machine (SVM), has been studied. This paper discussed how plant diseases could be detected via images of their leaves. This analysis seeks to validate a proposed system for an ap-propriate solution to the IoT-based environmental surveillance, water irrigation system management and an efficient approach for leaf disease detection on plants. The proposed multi hidden layers extreme machine classification delivers good performance of 99.12% in the classification of leaf diseases in comparison to the Support Vector Machine classification, which gives 98%.

Author 1: G. Balram

Author 2: K. Kiran Kumar

Keywords: Image acquisition; segmentation; feature extrac-tion; Internet of Things; plant disease

Paper 96: Learning Global Average Attention Pooling (GAAP) on Resnet50 Backbone for Person Re-identification Problem

Abstract: Person re-identification has been an extremely chal-lenging task in computer vision which has been seen as a success with deep learning approaches. Despite successful models, there are gaps in the form of unbalanced labels, poor resolution, uncertain bounding box annotations, occlusions, and unlabelled datasets. Previous methods applied deep learning approaches based on feature representation, metric learning, and ranking optimization. In this work, we propose Global Average Attention Pooling (GAAP) on Resnet50 applied on four benchmark Re-ID datasets for classification tasks. We also perform an extensive evaluation on the proposed Attention module with different deep learning pipelines as backbone architecture. The four benchmark person Re-ID datasets used is Market-1501, RAiD, Partial-iLIDS, and RPIfield. We computed cumulative matching characteristics (CMC) and mean Average Precision (mAP) as the performance evaluation parameters of the proposed against the state of the art. The results obtained have shown that the added attention layer has improved the overall recognition precision over the baselines.

Author 1: Syamala Kanchimani

Author 2: Maloji Suman

Author 3: P. V. V. Kishore

Keywords: Person re-identification; attention network; ResNet50; global average attention

Paper 97: Towards a Richer IndoWordNet with New Additions for Hindi and Gujarati Languages

Abstract: The authors of this research paper present a mech-anism for dealing with loanwords, missing words, and newly de-veloped terms inclusion issues in WordNets. WordNet has evolved as one of the most prominent Natural Language Processing (NLP) toolkits. This mechanism can be used to improve the WordNet of any language. The authors chose to work with the Hindi and Gujarati languages in this research work to achieve a higher quality research aspect because these are the languages with major dialects. The research work used more than 5000 Hindi verse-based data corpus instead of a prose-based data corpus.As a result, nearly 14000 Hindi words were discovered that were not present in the popular Hindi IndoWordNet, accounting for 13.23 percent of the total existing word count of 105000+. Working with idioms was a distinct method for the Gujarati language. Around 3500 idioms data were used, and nearly 900 Gujarati terms were discovered that did not exist in the IndoWordNet, accounting for nearly 1.4 percent of the total of 64000+ Gujarati words in the IndoWordNet. It will also contribute almost 14000 Hindi words and around 900 Gujarati words to the IndoWordNet project.

Author 1: Milind Kumar Audichya

Author 2: Jatinderkumar R. Saini

Author 3: Jatin C. Modh

Keywords: Gujarati; hindi; indian language WordNet; In-doWordNet; loanwords; WordNet

Paper 98: Prediction of Diabetic Retinopathy using Convolutional Neural Networks

Abstract: Diabetic retinopathy (DR) is among the most dan-gerous diabetic complications that can lead to lifelong blindness if left untreated. One of the essential difficulties in DR is early discovery, which is crucial for therapy progress. The accurate diagnosis of the DR stage is famously complicated and demands a skilled analysis by the expert being of fundus images. This paper detects DR and classifies its stage using retina images by applying conventional neural networks and transfer learning models. Three deep learning models were investigated: trained from scratch CNN and pre-trained InceptionV3 and Efficient-NetsB5. Experiment results show that the proposed CNN model outperformed the pre-trained models with a 9 to 25% relative improvement in F1-score compared to pre-trained InceptionV3 and EfficientNetsB5, respectively.

Author 1: Manal Alsuwat

Author 2: Hana Alalawi

Author 3: Shema Alhazmi

Author 4: Sarah Al-Shareef

Keywords: CNN; convolutional neural networks; deep learn-ing; transfer learning; medical imaging; diabetic retinopathy; retina fundus images

Paper 99: Real-time Egyptian License Plate Detection and Recognition using YOLO

Abstract: Automatic License Plate Detection and Recognition (ALPR) is one of the most significant technologies in intelligent transportation and surveillance across the world. It has many challenges because it affects by many parameters such as the country’s layout, colors, language, fonts, and several environmen-tal conditions so, there isn’t a consolidated ALPR system for all countries. Many ALPR methods have been proposed based on traditional image processing and machine learning algorithms since there aren’t enough datasets, particularly in the Arabic language. In this paper, we proposed a real-time ALPR system for the Egyptian license plate (LP) detection and recognition using Tiny-YOLOV3. It consists of two deep convolutional neural networks. The experimental results in the first available publicly Egyptian Automatic License Plate (EALPR) dataset show the proposed system is more robust in detecting and recognizing the Egyptian license plates and gives mean average precision values of 97.89% and 92.46% for LP detection and character recognition, respectively.

Author 1: Ahmed Ramadan Youssef

Author 2: Abdelmgeid Ameen Ali

Author 3: Fawzya Ramadan Sayed

Keywords: Automatic license plate recognition; Egyptian li-cense plate; Tiny-YOLOV3; CNN; eALPR dataset

Paper 100: Building an Arabic Dialectal Diagnostic Dataset for Healthcare

Abstract: Accurate diagnosis of patient conditions becomes challenging for medical practitioners in urban metropolitan cities. A variety of languages and spoken dialects impedes the diagnosis achieved through the exploratory journey a medical practitioner and patient go through. Natural language processing has been used in well-known applications, such as Google Translate, as a solution to reduce language barriers. Languages typically encountered in these applications provide the most commonly known, used or standardized dialect. The Arabic language can benefit from the common dialect, which is available in such applications. However, given the diversity of dialects in Arabic in the healthcare domain, there is a risk associated with incorrect interpretation of a dialect, which can impact the diagnosis or treatment of patients. Arabic language dialect corpuses published in recent research work can be applied to rule-based natural language applications. Our study aims to develop an approach to support medical practitioners by ensuring that the diagnosis is not impeded based on the misinterpretation of patient responses. Our initial approach reported in this work adopts the methods used by practitioners in the diagnosis carried out within the scope of the Emirati and Egyptian Arabic dialects. In this paper, we develop and provide a public Arabic Dialect Dataset (ADD), which is a corpus of audio samples related to healthcare. In order to train machine learning models, the dataset development is designed with multi-class labelling. Our work indicates that there is a clear risk of bias in datasets, which may come about when a large number of classes do not have enough training samples. Our crowd sourcing solution presented in this work may be an approach to overcome the sourcing of audio samples. Models trained with this dataset may be used to support the diagnosis made by medical practitioners.

Author 1: Jinane Mounsef

Author 2: Maheen Hasib

Author 3: Ali Raza

Keywords: Dialectal Arabic (DA); healthcare diagnosis; natu-ral language processing (NLP); multi-class labeling; crowd sourc-ing

Paper 101: Design of Higher-Dimensional Hyperchaotic System based on Combined Control and its Encryption Application

Abstract: According to the anti-control principle of chaos, a combined control method is proposed based on a class of asymptotically stable linear systems with multiple controllers. A higher-dimensional hyperchaotic system is investigated by the Lyapunov exponents method and equilibrium points analysis, and it exists the largest number of positive Lyapunov exponents. The chaotic pseudo-random sequences of the higher-dimensional hyperchaotic system can pass all NIST tests after preprocess-ing, and behave better chaotic characteristics. Meanwhile, a new encryption algorithm of image information with position scrambling, sequential diffusion and reverse diffusion is designed based on the chaotic pseudo-random sequences. The experiments of image information are given to verify the effectiveness and feasibility of the encryption algorithm. Finally, the security analyses are also discussed by the key sensitivity, differential attack and statistical analysis. It is shown that the encryption algorithm has large enough key space and can be applied to secure communication.

Author 1: Kun Zhao

Author 2: Jianbin He

Keywords: Hyperchaotic system; positive lyapunov exponent; chaotic pseudo-random sequence; image encryption

Paper 102: Exploring Regression-based Approach for Sound Event Detection in Noisy Environments

Abstract: Sound-event detection enables machines to detect when a particular sound event has occurred in addition to classifying the type of event. Successful detection of various sound events is paramount in building secure surveillance systems and other smart home appliances. However, noisy events and environ-ments exacerbate the performance of many sound event detection models, rendering them ineffective in real-world scenarios. Hence, the need for robust sound event detection algorithms in noisy environments with low inference times arises. You Only Hear Once (YOHO) is a purely convolutional architecture that uses a regression-based approach for sound-event-detection instead of the more common, frame-wise classification-based approach. The YOHO architecture proved robust in noisy environments, outperforming convolutional recurrent neural networks popular in sound event detection systems. Additionally, different ways to enhance the performance of the YOHO architecture are explored, experimenting with different computer vision architectures, dy-namic convolutional layers, pretrained audio neural networks and data augmentation methods to help improve the performance of the models on noisy data. Amongst several modifications to the YOHO architecture, the Frequency Dynamic Convolution Layers helped improve the internal model data representations by enforcing frequency-dependent convolution operations, which helped improve YOHO performance on noisy audios in outdoor and vehicular environments. Similarly, the FilterAugment data augmentation method and Convolutional Block Attention Module helped improve YOHO’s performance on the VOICe dataset containing noisy audios by augmenting the data and improving internal model representations of the input audio data using attention, respectively.

Author 1: Soham Dinesh Tiwari

Author 2: Karanth Shyam Subraya

Keywords: Sound Event Detection (SED); sound event clas-sification; frequency dynamic convolution; audio processing; Fil-terAugment; data augmentation; vision transformers; Pretrained Audio Neural Networks (PANN); Convolutional Block Attention Module (CBAM)

Paper 103: An Efficient Parallel Algorithm for Clustering Big Data based on the Spark Framework

Abstract: The principal objective of this paper is to provide a parallel implementation focused on the main steps of the parameter-free clustering algorithm based on K-means (PFK-means) using the Spark framework and a machine learning-based model to process Big Data. Thus, the process consists of parallelizing the main tasks of the first stage of the PFK-means clustering algorithm using successive RDD functions. Then, the parallel K-means provided by Spark MLlib is invoked by setting the cluster centers and the number of clusters determined in the previous step as input parameters of the parallel K-means. Furthermore, a comparison between the parallel designed algorithm and the parallel K-means was conducted using UCI data sets in terms of the sum of squared errors and the processing time. The experimental results, performed locally using the Spark framework, demonstrate the efficiency of the proposed solution.

Author 1: Zineb Dafir

Author 2: Said Slaoui

Keywords: Clustering; big data; spark; parallel computing; parallel K-means

Paper 104: Development of a Low-Cost Teleoperated Explorer Robot (TXRob)

Abstract: Natural disasters such as earthquakes or mudslides destroy everything in their path, causing buildings to collapse, which can cause people to lose their lives or suffer permanent injuries. Rescuers and firefighters are responsible for entering these ruined buildings, this work being very dangerous for them because they can get trapped in the rubble or suffocate due to the harmful gases found inside these buildings. Taking into consideration the risk in this type of operations, technological innovations can be used to help in the exploration of ruined build-ings and the rescue of people. Therefore, this article describes the development of TXRob, a low-cost teleoperated robot used in the exploration of post-disaster scenarios. TXRob has artificial vision, environmental gas recognition sensors, a real-time data display panel, is sized to enter buildings, and is capable of moving over uneven surfaces, such as debris or cracks, thanks to its track system. A human operator can remotely monitor and control the robot. The TXRob’s versatility as well as sensors performance has been tested on uneven and harsh surfaces in a simulated disaster environment. These tests suggest that the designed robot is suitable for use in rescue situations.

Author 1: Rafael Verano M

Author 2: Jose Caceres S

Author 3: Abel Arenas H

Author 4: Andres Montoya A

Author 5: Joseph Guevara M

Author 6: Jarelh Galdos B

Author 7: Jesus Talavera S

Keywords: Rescue robot; teleoperation; low cost robot; artificial vision

Paper 105: Automatic Tariff Classification System using Deep Learning

Abstract: The tariff fraction is the universal form of identi-fying a product. It is very useful because it helps to know the tariff that the product must pay when entering or leaving the country, in this case Mexico. Coffee is a complicated product to identify correctly due to its variants, which at first glance are not distinguishable, which can cause confusion and the tariff to be charged incorrectly. Therefore, the main objective of this project was to develop a system based on Deep Learning models, which allow to identify the tariff code of coffee to import or export this product through the analysis of digital images in real time, generating automatically a general report with this information for the customs broker. The developed system allows speeding up the process of assigning the tariff fraction, and also allows the correct assignment of the tariff fraction, avoiding confusion with other products and the wrong collection of the tariff. It is important to mention that the system, although for the moment it is focused on the country of Mexico, can be used in all customs offices since the tariff fraction is universal. The evaluation of the models was carried out with cross-validation, obtaining an effectiveness of more than 80%, and the tariff fraction assignment model had an effectiveness of 90%.

Author 1: German Cuaya-Simbro

Author 2: Irving Hernandez-Vera

Author 3: Elias Ruiz

Author 4: Karina Gutierrez-Fragoso

Keywords: Machine learning; digital image processing; automation process

Paper 106: Fake News Detection in Social Media based on Multi-Modal Multi-Task Learning

Abstract: The popularity of social media has led to a substantial increase of data. The task of fake news detection is very important, because the authenticity of posts cannot be guaranteed. In recent years, fake news detection combining multimodal information such as images and videos has attracted wide attention from scholars. However, the majority of research work only focuses on the fusion of multi-modal information, while neglecting the role of external evidences. To address this challenge, this paper proposes a fake news detection method based on multi-modal and multi-task learning. When learning the representation of the news posts, this paper models the interaction between images and texts in posts and external evidences through a multi-level attention mechanism, and uses evidence veracity classification as an auxiliary task, so as to improve the task of fake news detection. Authors conduct comprehensive experiments on a public dataset, and demonstrate that the proposed method outperforms several state-of-the-art baselines. The ablation experiment proves the effectiveness of the auxiliary task of evidence veracity in fake news detection.

Author 1: Xinyu Cui

Author 2: Yang Li

Keywords: Multi-modal fake news; multi-task learning; external evidences; multi-level attention mechanism

Paper 107: A Model Driven Approach for Unifying user Interfaces Development

Abstract: In this paper, we dealt with the rapid development of client web applications (frontend) in a context where development frameworks are legion. In effect with the digital transformation due to the COVID-19 pandemic we are witnessing an ever-increasing demand of the application development in a relatively short time. To this is added the lack of skilled developers on constantly evolving technologies. We therefore offer a low-code platform for the automatic generation of client web applications, regardless of the platform or framework chosen. First, we defined an interface design methodology based on a portal. We then implemented our model driven architecture which consisted of defining a modeling and templating language, centered on user data, flexible enough to not only be used in various fields but also be easily used by a citizen developer.

Author 1: Henoc Soude

Author 2: Kefil Koussonda

Keywords: Model driven; user interface; modeling language; templating language; low code; citizen developer

Paper 108: Acceptance of YouTube for Islamic Information Acquisition: A Multi-group Analysis of Students’Academic Discipline

Abstract: YouTube has been recognized as an important information source for the millennial generation. This paper aims to identify the factors affecting Malaysian higher education students’ acceptance of YouTube for Islamic information acquisition and to investigate if any notable distinction that exists between the students’ path coefficients in Islamic academic discipline and the other disciplines. Employing the Unified Theory of Acceptance and Use of Technology model (UTAUT) as its theoretical foundation, data were collected by distributing a self-administered survey to 795 students actively using YouTube for information seeking. Partial least squares structural equation modelling (PLS-SEM) and multi-group analysis (MGA) in SmartPLS 3.2.7 software were used to analyze the data. Three constructs of the UTAUT model, performance expectancy, effort expectancy and social influences, were found to significantly and positively influence behavioural intention to use YouTube for Islamic information acquisition in both groups of students. Facilitating conditions demonstrates significantly negative relationship with YouTube acceptance for students in other academic disciplines than for Islamic academic discipline. Additionally, the MGA analysis’ findings suggest that determinants’ factor coefficients of YouTube acceptance for Islamic information acquisition are not significantly different between students in Islamic academic disciplines and the other disciplines. This study validates the UTAUT model to understand the determinant of social media application usage in a new study context.

Author 1: M. S Ishak

Author 2: A Sarkowi

Author 3: M. F Mustaffa

Author 4: R Mustapha

Keywords: Unified Theory of Acceptance and Use of Technology (UTAUT); YouTube; information acquisition; student knowledge; Partial Least Square