Abnormal data detection of power marketing audit based on DSR-Net

Abstract

With the gradual realization of the goal of "full coverage, full collection and full cost control" of the power consumption information acquisition system established by the State Grid Corporation of China, the smart grid with the basic technical characteristics of informatization, automation and interaction has entered the stage of full completion. Therefore, the application of power big data technology has become an inevitable requirement for the intelligent development of the power industry, and power data anomaly detection has always been the top priority. Compared with traditional methods, this method is more able to collect the same feature relationship between data, and has strong generalization ability. After experimental comparison, the F1 score evaluation index of DSR-Net is 92.56, which is 9.33% higher than the traditional support vector machine algorithm.