A Comparative Study of Photovoltaic Power Prediction Models Based on BP and XGBoost

Photovoltaic (PV) power generation is influenced by fluctuating environmental factors and exhibits significant non-linearity, making accurate prediction challenging. This paper constructs photovoltaic power prediction models based on BP neural network and XGBoost algorithms, and experimentally validates their predictive performance. The models are evaluated using metrics such as root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>). The experimental results indicate that the BP neural network demonstrates stable performance in handling complex nonlinear relationships, while XGBoost achieves higher prediction accuracy and better generalization ability when the dataset is relatively small. This study provides a reference for the selection of photovoltaic power prediction models and lays the foundation for improving prediction accuracy in renewable energy power systems.