Comparison of Artificial Neural Network and Long Short-Term Memory for Modelling Crude Palm Oil Production in Indonesia

Indonesia is one of the largest producers and exporters of Crude Palm Oil (CPO), making CPO production crucial to the country's economic stability. Accurate forecasting of CPO production is essential for effective inventory management, export-import strategy, and economic planning. Traditional time series methods like ARIMA have limitations in modeling nonlinear data, leading to the adoption of machine learning approaches such as Artificial Neural Network (ANN) and Long Short-Term Memory (LSTM). This study compares the performance of ANN, a general neural network, and LSTM, a neural network specifically designed for time series data, in predicting CPO production in Indonesia. Data from 2003 to 2022 were used to train and evaluate both models with various hyperparameter tuning configurations. The results indicate that while both models provide excellent forecasting accuracy, with MAPE values below 10%, the LSTM model achieved a lower out-of-sample MAPE of 5.78% compared to ANN’s 6.87%, suggesting superior performance by LSTM in capturing seasonal patterns in CPO production. Consequently, LSTM is recommended as the preferred model for CPO production forecasting due to its enhanced ability to handle temporal dependencies and nonlinear patterns in the data.

@article{Ulama2025,
  title     = {Comparison of Artificial Neural Network and Long Short-Term Memory for Modelling Crude Palm Oil Production in Indonesia},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {16},
  number    = {1},
  year      = {2025},
  publisher = {The Science and Information Organization},
  author    = {Brodjol Sutijo Suprih Ulama and Robi Ardana Putra and Fausania Hibatullah and Mochammad Reza Habibi and Mochammad Abdillah Nafis},
  doi       = {10.14569/IJACSA.2025.0160171},
  url       = {https://doi.org/10.14569/IJACSA.2025.0160171}
}