Performance Evaluation of Monthly Electricity Demand Forecasting in Thailand Using Machine Learning and Statistical Models

Teerapong Chongwat; Artitayaporn Rojarath

doi:10.69650/rast.2026.265801

Authors

Teerapong Chongwat Faculty of Informatics, Mahasarakham University, Mahasarakham 44150, Thailand
Artitayaporn Rojarath Multi-agent Intelligent Simulation Laboratory (MISL) Research Unit, Department of Information Technology, Faculty of Informatics, Mahasarakham University, Mahasarakham 44150, Thailand

DOI:

https://doi.org/10.69650/rast.2026.265801

Keywords:

Artificial Neural Networks, Autoregressive Integrated Moving Average, Electricity Consumption Forecasting, Long Short-Term Memory, Time Series Analysis

Abstract

This research aims to systematically evaluate and compare the performance of time series forecasting models for predicting monthly electricity consumption in Thailand. The study focuses on assessing model suitability and the impact of parameter configurations on forecasting accuracy. Three forecasting approaches are considered, including an Artificial Neural Network (ANN), a Long Short-Term Memory (LSTM) network, and an AutoRegressive Integrated Moving Average (ARIMA) model. All models are trained and evaluated using the same dataset to ensure a fair comparison. The experimental results demonstrate that the ANN model consistently achieves the best overall performance when evaluated across all electricity consumer groups. The ANN model attains
a Mean Absolute Error (MAE) of 71.72, a Mean Squared Error (MSE) of 20,050.34, a Root Mean Squared Error (RMSE) of 93.06, a Mean Absolute Percentage Error (MAPE) of 9.54%, and an R² value of 0.45, outperforming both the LSTM and ARIMA models in most evaluation metrics.
In contrast, the ARIMA model demonstrates higher prediction errors in several consumer groups, indicating challenges in capturing nonlinear patterns and demand variability. The forecasting results of the ANN model indicate that the business and industrial sector constitutes the group with the highest level of electricity consumption.

References

Uddin, G. S., Hasan, M. B., Phoumin, H., Taghizadeh-Hesary, F., Ahmed, A. and Troster, V., Exploring the critical demand drivers of electricity consumption in Thailand. Energy Economics. 125 (2023) 106875, doi: https://doi.org/10.1016/j.eneco.2023.106875.

Gholamnia, M., Eslamirad, N., Sajadi, P., Masoumi, S., Shahabi, H. and Pilla, F., Dynamic electricity pricing model with hourly and monthly adjustments: A time series-based approach. Energy Reports. 13 (2025) 5238-5251, doi: https://doi.org/10.1016/j.egyr.2025.04.058.

Cui, Z., Wu, J., Lian, W. and Wang, Y.-G., A novel deep learning framework with a COVID-19 adjustment for electricity demand forecasting. Energy Reports. 9 (2023) 1887–1895, doi: https://doi.org/10.1016/j.egyr.2023.01.019.

Alsamraee, S. A. and Khanna, S., Long-term electricity demand forecasting of a university campus based on advanced deep learning artificial neural network algorithms. Results in Engineering. 27 (2025) 106683, doi: https://doi.org/10.1016/j.rineng.2025.106683.

Su, Z., Zhang, J., Yang, Z. and Ma, L., A hybrid monthly electricity demand forecasting model combining an Hodrick-Prescott filter, recurrent neural networks, and autoregressive integrated moving average. Energy and AI. 22 (2025) 100600, doi: https://doi.org/10.1016/j.egyai.2025.100600.

Sobot, T., Stankovic, V. and Stankovic, L., Human in the loop active learning for time-series electrical measurement data. Engineering Applications of Artificial Intelligence. 133 (2024) 108589, doi: https://doi.org/10.1016/j.engappai.2024.108589.

Backhaus, F., Brucke, K., Ruckdeschel, P. and Schlüters, S., e-values based continuous-time model selection for residential electricity demand forecasts. Energy Build. 349 (2025) 116452, doi: https://doi.org/10.1016/j.enbuild.2025.116452.

Ghimire, S., Deo, R. C., Casillas-Pérez, D. and Salcedo-Sanz, S., Efficient daily electricity demand prediction with hybrid deep-learning multi-algorithm approach. Energy Conversion and Management. 297 (2023) 117707, doi: https://doi.org/10.1016/j.enconman.2023.117707.

Wang, W., Shimakawa, H., Jie, B., Sato, M. and Kumada, A., BE-LSTM: An LSTM-based framework for feature selection and building electricity consumption prediction on small datasets. Journal of Building Engineering. 102 (2025) 111910, doi: https://doi.org/10.1016/j.jobe.2025.111910.

Huuki, H., Ruokamo, E., Kopsakangas-Savolainen, M., Belonogova, N., Sridhar, A. and Honkapuro, S., House and socio-demographic features vs. electricity consumption time series in main heating mode classification. The Electricity Journal. 37 (2024) 107373, doi: https://doi.org/10.1016/j.tej.2024.107373.

Hussein, A. and Awad, M., Time series forecasting of electricity consumption using hybrid model of recurrent neural networks and genetic algorithms. Measurement: Energy. 2 (2024) 100004, doi: https://doi.org/10.1016/j.meaene.2024.100004.

Pérez-Chacón, R., Asencio-Cortés, G., Troncoso, A. and Martínez-Álvarez, F., Pattern sequence-based algorithm for multivariate big data time series forecasting: Application to electricity consumption. Future Generation Computer Systems. 154 (2024) 397–412, doi: https://doi.org/10.1016/j.future.2023.12.021.

Lazcano, A. and Jaramillo-Morán, M. A., Data preprocessing techniques and neural networks for trended time series forecasting. Applied Soft Computing. 174 (2025) 113063, doi: https://doi.org/10.1016/j.asoc.2025.113063.

Dercole, F., Sangiorgio, M. and Schmirander, Y., An empirical assessment of the universality of ANNs to predict oscillatory time series. IFAC-PapersOnLine. 53 (2020) 1255–1260, doi: https://doi.org/10.1016/j.ifacol.2020.12.1850.

Kim, M. K., Kim, Y.-S., Fu, N., Liu, J., Wang, J., Lee, S. and Srebric, J., Advanced techniques for electricity consumption prediction in buildings using comparative correlation analysis, data normalization, and Long Short-Term Memory (LSTM) networks: A case study of a U.S. commercial building. Energy Reports. 14 (2025) 56-65, doi: https://doi.org/10.1016/j.egyr.2025.05.074.

Hassanpouri Baesmat, K., Shokoohi, F. and Farrokhi, Z., SP-RF-ARIMA: A sparse random forest and ARIMA hybrid model for electric load forecasting. Global Energy Interconnection. 8 (2025) 486–496, doi: https://doi.org/10.1016/j.gloei.2025.04.003.

Gao, F. and Shao, X., Electricity consumption prediction based on a dynamic decomposition-denoising-ensemble approach. Engineering Applications of Artificial Intelligence. 133 (2024) 108521, doi: https://doi.org/10.1016/j.engappai.2024.108521.

Jamil, R., Hydroelectricity consumption forecast for Pakistan using ARIMA modeling and supply-demand analysis for the year 2030. Renewable Energy. 154 (2020) 1-10, doi: https://doi.org/10.1016/j.renene.2020.02.117.

Manohar, B., Das, R. and Lakshmi, M., A hybridized LSTM-ANN-RSA based deep learning models for prediction of COVID-19 cases in Eastern European countries. Expert Systems with Applications. 256 (2024) 124977, doi: https://doi.org/10.1016/j.eswa.2024.124977.

Silva, D. G. da and Meneses, A. A. de M., Comparing long short-term memory (LSTM) and bidirectional LSTM deep neural networks for power consumption prediction. Energy Reports. 10 (2023) 3315–3334, doi: https://doi.org/10.1016/j.egyr.2023.09.175.

Srisurin, K., Tanwattanapong, S., Khaikratok, T. and Chinnayomphanit, A. Forecasting Model in Electricity Consumption in Thailand. in The 13th Hatyai National and International Conference. (2022), 607–622.

Meftah, E. and Adel, M., A Comparative Analysis of the ARIMA and LSTM Predictive Models and Their Effectiveness for Predicting Wind Speed. Energies. 14 (2021) 6782, doi: https://doi.org/10.3390/en14206782.

Srisurin, K., Tanwattanapong, S., and Sompongtham, C. Forecasting of Peak Demand in Thailand. in the 10th National and the 8th International Conference on Research and Innovation. (2023), 591–606.

Promnuchanont, T., Kosarat, R. and Jaidee, W., Comparative Analysis of Time Series Forecasting Models for Predicting Tourist Arrivals in Chiang Mai. KKU Science Journal. 52 (2024) 289 - 302, doi: https://doi.org/10.14456/kkuscij.2024.23.

Anekkunwat, P. Forecasting international tourist arrivals to Thailand using time series and artificial neuron network. Master of Science, Thesis, Chulalongkorn University, Bangkok, Thailand, (2020).

Tarmanini, C., Sarma, N., Gezegin, C. and Ozgonenel, O., Short term load forecasting based on ARIMA and ANN approaches. Energy Reports. 9 (2023) 550–557, doi: https://doi.org/10.1016/j.egyr.2023.01.060.

Corradini, F., Gerosa, F., Gori, M., Lucheroni, C., Piangerelli, M. and Zannotti, M., A systematic literature review of spatio-temporal graph neural network models for time series forecasting and classification. Neural Networks. 195 (2026) 108269, doi: https://doi.org/10.1016/j.neunet.2025.108269.

Zhang, W., Zhan, H., Sun, H. and Yang, M., Probabilistic load forecasting for integrated energy systems based on quantile regression patch time series Transformer. Energy Reports. 13 (2025) 303–317, doi: https://doi.org/10.1016/j.egyr.2024.11.057.