Optimization of energy usage decisions through predictive modeling of consumption patterns

Accurate energy consumption prediction is a critical resource allocation, reducing operational costs, and supporting sustainable energy management. This study presents a machine learning-based energy consumption prediction system, comparing the performance of Linear Regression, Ridge Regression, Lasso Regression, and Decision tree models. Using a historical dataset of consumer energy usage, key features were extracted and preprocessed to enhance predictive accuracy. Model evaluation employed Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Coefficient of Determination metrics. Results indicate that while Linear Regression offered competitive baseline performance, Ridge and Lasso Regressions achieved improved stability against multicollinearity, and the Decision Tree Model delivered the highest predictive accuracy, albeit with potential overfitting risks. The findings contribute to developing scalable, data-driven forecasting frameworks applicable to smart grids, policy planning, and demand-side management.