Machine Learning-based Electrical Energy Consumption Prediction

Electrical energy is a fundamental resource widely used in both daily life and various industrial sectors. Accurate prediction of electricity consumption is critical for managing electrical energy. In this paper, we propose a machine learning model to predict electricity consumption with data from the City of Chicago in Illinois. We modeled and validated five models: Linear Regression, Decision Tree Regression, Random Forest Regression, XGBoost Regression(eXtreme Gradient Boosting Regression), and LightGBM Regression(Light Gradient Boosting Machine Regression). The LightGBM Regression has shown the most suitable model for predicting electricity consumption. We also visualized the effects of independent variables on electricity consumption with XAI(eXplainable Artificial Intelligence), SHAP(SHapley Additive exPlanations). We expect that our proposed model helps predicting and managing electrical energy effectively.