Forecasting the Solar Panel Output Using the Machine Learning Methods

Importance of accurate analysis and forecasting the power system performance is constantly growing. Efficient use of data related to the solar panels performance makes it possible not only to optimize their operation, but also to develop more accurate forecasting models based on the machine learning. One of the key steps in this process is collection and analysis of the variety of data reflecting both the system environmental characteristics and technical parameters. This study presents a dataset containing solar energy parameters such as temperature, humidity, wind speed, and visibility, collected from meteorological sources. The dataset is analyzed using two machine learning models, Random Forest Regressor (RFR) and Multi-Layer Perceptron (MLP), to predict solar energy output. The models were trained using historical meteorological data, and their performance was evaluated based on Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and the Coefficient of Determination (R2). Feature selection was performed using correlation analysis to identify the most influential parameters affecting solar energy prediction. Experimental results demonstrate that RFR outperforms MLP in terms of predictive accuracy, with lower MAE and RMSE values and a higher R2 score. These findings highlight the potential of ensemble-based models for enhancing solar energy forecasting and optimizing energy management strategies.