Abstract

The use of several machine learning regression models for the prediction of power costs is explored in this work. The following models are taken into consideration: Gradient Boosting, Adaboosting, Lgbmregressor, Cataboost regressor, Random Forest Regressor, XGBoost Regressor, Support Vector Regressor, and Stacking Regressor. Historical power pricing data with variables like time, demand, and weather make up the training and assessment dataset. A portion of the dataset is used to train each model, which is then adjusted to maximize prediction performance. To evaluate each model's precision, resilience, and computational effectiveness, a comparative study is done. Insights into the applicability of various regression models for the forecast of power prices are provided by the study's findings, which may help decision-makers optimize the planning and management of energy resources.