A Novel Ensemble Model for Load Forecasting: Integrating Random Forest, XGBoost, and Seasonal Naive Methods

As energy systems become more complex due to the integration of renewable sources and varying consumption patterns, current load forecasting methods often fall short in accuracy and adaptability, leading to inefficient resource allocation and potential grid instability. To address these challenges, this study introduces a hybrid ensemble model that elevates predictive accuracy by integrating multiple preprocessing techniques and algorithms, including Random Forest, XGBoost, and Seasonal Naive methods. The model is trained under both Standardized and Sparse Normalized data and is fused by assigning different weights. Compared to standalone models, the hybrid approach excels across multiple performance metrics. Future research may delve into the model’s applicability for real-time data processing for instantaneous decision-making in energy management.