Power outage classification prediction method based on bagging ensemble learning

Power outage prediction can provide reference for power companies' power outage decision-making, reduce frequency of power outages on the user side, and improve reliability of low-voltage power supply. In this paper, Bagging Ensemble model is used to predict the power failure problem. First, paper construct spatial location matrix of different regions according to geographic location relationship of different regions, and construct spatial features through QR matrix decomposition. Then, bagging ensemble learning framework is used to randomly resample the data, train different base classifiers, and integrate classifiers according to the combination strategy. Finally, paper use integrated learning model to make predictions on the data. The common classification model evaluation index and AOC curve are used to measure performance of model. Experimental results show prediction model proposed in this paper has good performance for classification prediction of outage data.