Machine Learning-Based Solar Panel System Failure Detection Using Random Forest, Gradient Boosting, and AdaBoost

This research applies a combination of machine learning methods - Random Forest, Gradient Boosting, and AdaBoost - to detect solar panel system failures using historical data. The ensemble model developed achieved an accuracy rate of 98.17 percent, with the best performance shown in Class 1 which achieved perfect scores (Precision, Recall, and F1-Score = 1.0000). Meanwhile, Class 0 and Class 2 also show excellent performance with Precision, Recall and F1-Score values above 0.96. These results prove that combining algorithms in an ensemble model produces higher detection accuracy than using a single algorithm, and can increase operational efficiency and maintenance of solar panel systems.