Energy Personalized Recommender for the PowerEye Home Energy Management System

The Energy Personalized Recommender (EPR) is a novel backend module designed to optimize energy management within home energy management systems (HEMSs). Existing research explores various energy optimization strategies but is often constrained by synthetic datasets, simulation-based validation, or limited personalization. To address these limitations, EPR is designed for real-world deployment and personalization. By employing machine learning algorithms, the EPR analyzes appliance-level energy consumption in real time and integrates various algorithms to enhance user interaction with energy usage data. These algorithms include the Energy Goal Algorithm, which allows users to set and monitor monthly energy cost goals, and the Peak Time Algorithm, which advises on shifting usage to off-peak times. The Phantom Mode Algorithm identifies low-power or standby appliances, promoting energy conservation, while the Energy Baseline Algorithm predicts hourly energy usage to detect deviations. In evaluating the system, k-means clustering for phantom appliances achieved silhouette scores of 0.478 to 0.954, indicating reliable performance. Baseline prediction models showed RMSE values between 0.003 and 1.526, depending on the appliance. A single-day evaluation involving three participants showed mixed consumption changes: two participants recorded decreases, while the third recorded an increase, resulting in a 19.74% aggregate increase. These preliminary findings provide a foundation for advancing personalized energy management and motivate longer controlled trials to determine efficient and sustained energy savings through EPR systems.

@article{Hejazi2026,
  title     = {Energy Personalized Recommender for the PowerEye Home Energy Management System},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {17},
  number    = {9},
  year      = {2026},
  publisher = {The Science and Information Organization},
  author    = {Reem Hejazi and Sofianiza Abd Malik and Anees Ara and Layla Alfawzan},
  doi       = {10.14569/IJACSA.2026.0170989},
  url       = {https://doi.org/10.14569/IJACSA.2026.0170989}
}