In response to the problem of a large amount of existing user electricity consumption behavior data and imbalanced positive and negative samples,resulting in low accuracy in electricity anomaly detection,this article uses the synthetic minority oversampling technique algorithm to add negative sample data and construct a dataset with relatively balanced positive and negative samples.Then,the classification and regression tree decision tree model is used to extract features from the rebalanced sample dataset,obtaining classification results and visualizing them.The algorithm is simulated and implemented through five fold cross validation to analyze the reliability and stability of the model.The experiment shows that the method proposed in this article effectively achieves abnormal detection of user electricity consumption behavior and improves recognition accuracy.
关键词
数据不均衡/决策树/用电异常检测
Key words
data imbalance/decision tree/abnormal electricity consumption detection
维普
万方数据