MLP Sliding-Window Forecasting for Electricity Load Prediction: A Multi-Scale Evaluation using an ONNX-based Java Framework

Farid Morsidi; Asma Hanee  Ariffin; Rohaizah Abdul Wahid

doi:10.24191/jaeds.v6i1.164

Authors

Farid Morsidi Computing Department, Faculty of Computing & Meta-Technology, Universiti Pendidikan Sultan Idris, Perak, Malaysia.
Asma Hanee Department of Software Engineering and Smart Technology, Universiti Pendidikan Sultan Idris, Perak, Malaysia.
Rohaizah Department of Software Engineering and Smart Technology, Universiti Pendidikan Sultan Idris, Perak, Malaysia

DOI:

https://doi.org/10.24191/jaeds.v6i1.164

Keywords:

Time-series forecasting, Deep learning, ONNX, Multi-layer perceptron, LSTM

Abstract

Deep learning models for time-series forecasting offer significant potential but face deployment challenges. This paper advances a modular framework for deploying ONNX models in Java, building on previous work. The research paper assesses the MLP sliding-window architecture through different time intervals which extends the previous LSTM-based system to establish primary performance standards that future research can use for comparison. The framework incorporates additional features for enhanced evaluation, including five metrics (MAE, RMSE, MAPE, SMAPE, R²), automated archiving, and high-quality graphical exports. Evaluations using electricity consumption data (ETTh1) and benchmark datasets show that the MLP sliding-window model outperforms LSTM in terms of R², MAE, and MAPE, achieving scores of 0.9895, 0.5907, and 7.19%, respectively, indicating high accuracy across various domain. The framework also implements a custom threshold (ε = 1e−10) to handle near-zero values in MAPE calculation, ensuring reliable result storage for reproducibility. The findings reveal that simpler MLP designs can either match or exceed LSTM performance in specific scenarios while offering faster processing and easier implementation. Detailed comparisons and guidelines for deployment are included, along with insights into model selection criteria

Downloads

Download data is not yet available.

References

Y. Wang, P. Jia, Z. Shu, K. Liu, and A. Rashid, “Multidimensional precipitation index prediction based on C NN -LSTM hybrid framework,” 2025, doi: https://doi.org/10.48550/arXiv.2504.20442.

M. Asif, M. M. Kuglitsch, I. Pelivan, and R. Albano, “Review and Intercomparison of Machine Learning Applications for Short-term Flood Forecasting,” Water Resour. Manag., pp. 1971–1991, 2025, doi: 10.1007/s11269-025-04093-x.

A. Poghosyan et al., “An enterprise time series forecasting system for cloud applications using transfer learning,” Sensors, vol. 21, no. 5, pp. 1–28, 2021, doi: 10.3390/s21051590.

F. Morsidi, “A Modular Java-Based Framework for Deploying ONNX Time-Series Forecasting Models : A Rainfall Prediction Case Study,” J. Appl. Eng. Des. Simul., vol. 5, no. 2, pp. 90–104, 2025, doi: 10.24191/jaeds.v5i2.144.

S. D. Latif et al., “Assessing rainfall prediction models: Exploring the advantages of machine learning and remote sensing approaches,” Alexandria Eng. J., vol. 82, no. May, pp. 16–25, 2023, doi: 10.1016/j.aej.2023.09.060.

P. Kanchan, “Rainfall Analysis and Forecasting Using Deep Learning Technique,” J. Informatics Electr. Electron. Eng., vol. 2, no. 2, pp. 1–11, 2021, doi: 10.54060/jieee/002.02.015.

D. Sun, J. Wu, H. Huang, R. Wang, F. Liang, and H. Xinhua, “Prediction of Short-time rainfall based on deep learning,” Math. Probl. Eng., vol. 2021, 2021, doi: 10.1155/2021/6664413.

A. Bhardwaj, “Time Series Forecasting with Recurrent Neural Networks: An In-depth Analysis Time Series Forecasting with Recurrent Neural Networks: An In-depth Analysis and Comparative Study,” EDU J. Int. Aff. Res. (EJIAR), ISS, vol. 2, no. 4, 2024, [Online]. Available: https://edupublications.com/index.php/ejiar/article/view/36

R. Rosly, M. Man, A. Ngah, and N. S. A. Manan, “Multi-classifier models to improve the accuracy of fish landing application,” Int. J. Adv. Technol. Eng. Explor., vol. 11, no. 111, pp. 145–159, 2024, doi: 10.19101/IJATEE.2023.10102060.

M. Koklu and I. A. Ozkan, “Multiclass classification of dry beans using computer vision and machine learning techniques,” Comput. Electron. Agric., vol. 174, p. 105507, 2020, doi: https://doi.org/10.1016/j.compag.2020.105507.

D. Endalie, G. Haile, and W. Taye, “Deep learning model for daily rainfall prediction: Case study of Jimma, Ethiopia,” Water Supply, vol. 22, no. 3, pp. 3448–3461, 2022, doi: 10.2166/WS.2021.391.

R. Taylor, “Machine Learning Techniques for Fish Breeding Decision Making,” 2023 Wellingt. Fac. Eng. Symp., pp. 1–12, 2023, [Online]. Available: https://ojs.victoria.ac.nz/wfes/article/view/8422

S. M. M. Islam, S. I. Bani, and R. Ghosh, “Content-based Fish Classification Using Combination of Machine Learning Methods,” Int. J. Inf. Technol. Comput. Sci., vol. 13, no. 1, pp. 62–68, 2021, doi: 10.5815/ijitcs.2021.01.05.

P. Han, C. Du, J. Chen, and X. Du, “Minimizing Monetary Costs for Deadline Constrained Workflows in Cloud Environments,” IEEE Access, vol. 8, pp. 25060–25074, 2020, doi: 10.1109/ACCESS.2020.2971351.

M. Bacevicius and A. Paulauskaite-Taraseviciene, “Machine Learning Algorithms for Raw and Unbalanced Intrusion Detection Data in a Multi-Class Classification Problem,” Appl. Sci., vol. 13, no. 12, 2023, doi: 10.3390/app13127328.

F. M. Javed Mehedi Shamrat et al., “LungNet22: A Fine-Tuned Model for Multiclass Classification and Prediction of Lung Disease Using X-ray Images,” J. Pers. Med., vol. 12, no. 5, 2022, doi: 10.3390/jpm12050680.

R. A. Adewoyin, P. Dueben, P. Watson, Y. He, and R. Dutta, “TRU-NET: A deep learning approach to high resolution prediction of rainfall,” Mach. Learn., vol. 110, no. 8, pp. 2035–2062, 2021, doi: 10.1007/s10994-021-06022-6.

I. K. Opara, U. L. Opara, J. A. Okolie, and O. A. Fawole, “Machine Learning Application in Horticulture and Prospects for Predicting Fresh Produce Losses and Waste: A Review,” Plants, vol. 13, no. 9, pp. 1–21, 2024, doi: 10.3390/plants13091200.

V. Atashi and H. T. Gorji, “Enhanced flood prediction using LSTM and climate parameters: Multi-station analysis of snowmelt-induced flooding in the Red River of the North,” J. Hydroinformatics, vol. 27, no. 2, pp. 245–260, 2025, doi: 10.2166/hydro.2025.236.

A. Fayyazbakhsh, T. Kienberger, and J. Vopava-Wrienz, “Comparative Analysis of Load Profile Forecasting: LSTM, SVR, and Ensemble Approaches for Singular and Cumulative Load Categories,” Smart Cities, vol. 8, no. 2, 2025, doi: 10.3390/smartcities8020065.

P. Hess and N. Boers, “Deep Learning for Improving Numerical Weather Prediction of Heavy Rainfall,” J. Adv. Model. Earth Syst., vol. 14, no. 3, pp. 1–11, 2022, doi: 10.1029/2021MS002765.

Deep Java Library (DJL), “ONNX runtime engine for DJL,” 2024.

AWS Machine Learning Blog, “Simplified MLOps with Deep Java Library and ONNX,” Jun. 2023.

Deep Java Library (DJL), “Time series forecasting extension,” 2024.

E. Salcedo, “Graph Learning-based Regional Heavy Rainfall Prediction Using Low-Cost Rain Gauges,” 2024, doi: 10.1109/LA-CCI62337.2024.10814868.

A. E. Yilmaz and H. Demirhan, “Weighted kappa measures for ordinal multi-class classification performance,” Appl. Soft Comput., vol. 134, p. 110020, 2023, doi: 10.1016/j.asoc.2023.110020.

T. Indravattana, “Chula Digital Collections Solving multi-depot open pickup and delivery problem with capacity and distance restrictions by ant colony optimization algorithm Solving Multi - depot Open Pickup and Delivery Problem with Capacity and Distance Restrictions by A,” 2023.