An Anomaly Detection Method for Power User Side Data Based on DBN Neural Network

Abnormal data may be caused by faults, equipment damage, external interference, malicious attacks, and other reasons. If not detected and processed in a timely manner, it may have a serious impact on the operation of the power system. Therefore, developing an effective method for detecting abnormal data on the power user side has important practical significance. This article proposes an anomaly detection method for power user side data based on deep belief networks (DBN). Firstly, preprocess the power user side data, including denoising and normalization, and data cleaning. Secondly, the features of user side abnormal data are extracted through discrete point transformation. Finally, construct a DBN network model and set appropriate network structure and parameters. Use a trained DBN network to perform anomaly detection on new power user side data. The experimental results show that the method proposed in this paper can effectively improve the accuracy of data anomaly detection.