Forecasting Long-Term Electricity Consumption in Saudi Arabia Based on Statistical and Machine Learning Algorithms to Enhance Electric Power Supply Management

This study aims to develop statistical and machine learning methodologies for forecasting yearly electricity consumption in Saudi Arabia. The novelty of this include (i) determining significant features that have a considerable influence on consumption, (ii) utilizing Bayesian optimization algorithm (BOA) enhance the model’s hyperparameters, (iii) hybridizing BOA with algorithms, viz., support vector regression (SVR) nonlinear autoregressive networks exogenous inputs (NARX), modeling individually long-term (iv) comparing their performances widely used classical time-series integrated moving average (ARIMAX) regard accuracy, computational efficiency, generalizability, (v) future validation. population, gross domestic product (GDP), imports, refined oil products were observed be total coefficient determination R2 values all developed models are >0.98, indicating an excellent fit historical data. However, among three proposed models, BOA–NARX has best performance, improving accuracy (root mean square error (RMSE)) by 71% 80% compared ARIMAX BOA–SVR respectively. overall results confirm higher reliability methods can power system operators more accurately forecast ensure sustainability electric energy. also provide guidance helpful insights researchers understanding crucial research, emerging trends, new developments energy studies.