Forecasting Daily Electricity Consumption in Thailand Using Regression, Artificial Neural Network, Support Vector Machine, and Hybrid Models

Abstract

This article involves forecasting daily electricity consumption in Thailand. Electricity consumption data are provided by the Electricity Generating Authority of Thailand, the leading power utility state enterprise under the Ministry of Energy. Five forecasting techniques, including multiple linear regression, artificial neural network (ANN), support vector machine, hybrid models, and ensemble models, are implemented. The article proposes a hyperparameter tuning technique, called sequential grid search, which is based on the widely used grid search, for ANN and hybrid models. Auxiliary variables and indicator variables that can improve the models’ forecasting performance are included. From the computational experiment, the hybrid model of a multiple regression model to forecast the expected daily consumption and ANNs from the sequential grid search to forecast the error term, along with additional indicator variables for some national holidays, provides the best mean absolution percentage error of 1.5664% on the test data set.

An Improved Hybrid Approach for Daily Electricity Peak Demand Forecasting during Disrupted Situations: A Case Study of COVID-19 Impact in Thailand

“…The comparative performance between the proposed model and prior studies is shown in Table 11. This comparative analysis is conducted with the other models that appear in the literature such as benchmark model 1 [53], LSTM [88], XGBoost [89], and SVM [90]. All comparative models follow the same structure, as outlined in prior research.…”

“…According to the result shown in Table 11, the proposed model not only surpasses all benchmark models but also outperforms all other comparative models in all three measures (i.e., MAPE, RMSE, and MAE) The comparative performance between the proposed model and prior studies is shown in Table 11. This comparative analysis is conducted with the other models that appear in the literature such as benchmark model 1 [53], LSTM [88], XGBoost [89], and SVM [90]. All comparative models follow the same structure, as outlined in prior research.…”

“…According to the result shown in Table 11, the proposed model not only surpasses all benchmark models but also outperforms all other comparative models in all three measures (i.e., MAPE, RMSE, and MAE). 4.63% 1501 1153 LSTM [88] 5.15% 1632 1266 XGBoost [89] 4.12% 1322 1015 SVM [90] 4.13% 1305 1014…”

Application of random forest based on semi-automatic parameter adjustment for optimization of anti-breast cancer drugs

“…Support vector machine (SVM) is a supervised machine learning that can deal with classification and regression problems (36), the basic idea is to find the optimal classification hyperplane that completely separates the two classes of samples in the original space in the linearly divisible case and to use kernel methods in the nonlinear case to solve problems that are nonlinear in low-dimensional space as linearly integrable problems in high-dimensional space (37). Delineating the hyperplane can be defined as a linear equation:…”

Fostering Energy Resilience in the Rural Thai Power System—A Case Study in Nakhon Phanom

“…Several works have evaluated electric appliance ownership and electricity consumption, and some generated load profiles, for Thai residents. However, they often do not differentiate between urban and rural residents [25][26][27][28], focus solely on urban households [29,30], industries [31], educational institutes [32,33] or the whole country [34], were conducted with a small sample size [7,14,[35][36][37], or a combination of all of the above. Thus, we aim to close this gap and apparent research need in the following sections.…”