Application of Statistical and Machine Learning Models in Vietnam’s Energy Consumption Demand Forecasting

Energy consumption demand forecasting plays a critical role in the planning and development of the nation’s energy security, which underpins the 8-year Power Development Plan (PDP8) and Vietnam’s ambitious Net-Zero 2050 commitment. However, this task becomes more difficult while challenges the big data environment is filled with a lot of noise and high fluctuation data. In order to deal with the problem, this paper using four distinct models which are Linear Regression, Holt’s (Additive), PSO-GM (1,1), and Support Vector Regression (SVR) to conduct a rigorous comparative analysis to identify the most accurate forecasting model. The performance evaluated by MAE, RMSE, MAPE indexes based on the Vietnam’s total primary energy demand data from 1986 to 2024. To check the accuracy of forecasting model, this study slits the length of data was into two period time, first time for the training data (1986- 2016) and next time for the testing data set (2017-2024). The results decisively identified that the Holt’s model achieving significantly outperforming all counterparts with the lowest error metrics (MAE = 89.33, RMSE = 99.50, and a MAPE of 7.19%). This model is strongly suggested to forecast the Vietnam’s energy demand in the period time 2025 to 2030. Based on this model, the Vietnam’s energy demand will reach 1528.08 TWh and 1882.55 TWh in 2025 and 2030, respectively. Furthermore, this study provides empirical evidence that simpler, well-chosen statistical models can surpass complex alternatives in small-sample scenarios, offering a reliable quantitative baseline for policymakers to navigate infrastructure development and decarbonization challenges.