Enhanced Monthly Load Forecasting With RapidMiner-Based Deep Learning

Precise electrical demand forecasting is crucial for maintaining the reliability of the electricity supply, particularly in large urban centers. This study developed an artificial intelligence model with the ability to forecast daily electricity load over several months with high accuracy. The proposed model was trained and validated using historical energy consumption and meteorological data in a case study carried out in Ho Chi Minh City, Vietnam. Unlike previous MATLAB studies, this study employed the RapidMiner program to reduce calculation time and give a visual framework. The mean absolute percent error (MAPE) was used to evaluate prediction performance, yielding a MAPE of 0.52%, compared to 1.1% for Decision Tree and 8.9% for Support Vector Machine. Testing demonstrated that the proposed Deep Learning model significantly outperformed the baseline models. By incorporating feature extraction and explainability techniques, the model achieved high sensitivity to fluctuations, as indicated by an R-squared (R²) value of 0.999. These results suggest that the model is practical for real-world applications and can assist in improving power system operation planning.