Quantification of Uncertainty in PV Generation Prediction through Machine Learning Regularization Techniques
Year: 2024
Type: article
Abstract: This study explores the integration of machine learning and uncertainty quantification to advance the prediction of PV generation.We evaluated four models using ANN, LSTM, and XGBoost to assess forecast accuracy and reliability.Results indicate that the LSTM combined with XGBoost (Case 4) consistently achieved superior performance, with the lowest CV(RMSE) values across varying conditions, including high humidity.This outcome suggests that LSTM's capability to handle temporal data, alongside XGBoost's nonlinear predictive strength, significantly enhances overall model performance.Additionally, this model's robustness points to an effective strategy for mitigating PV generation inherent intermittency and prediction uncertainty.
Source: 한국생활환경학회지
Authors: Sung-Hyup Hong, Kwang Ho Lee
Institution: Korea University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4404340864