Abnormal detection method for user registration information of electricity supplier based on improved Adaboost algorithm

Due to the high efficiency of online office, it has promoted the continuous development of network offices platforms in various industries. As the largest electricity-selling enterprise, China's power grid company has also launched some online electricity trading platforms. However, due to its short launch time, its network security issues are the focus of attention. Especially in China where there are numerous power users, anomalies of registration information often occur. Thus, this paper proposes a registration information anomaly detection method based on an improved Adaboost algorithm. Firstly, the registration information is quantified based on its completeness, abnormality, consistency, and stability. Secondly, the quantized registration information is detected and abnormal registration information is identified based on the BP neural network and the Adaboost algorithm. Finally, the feasibility of the method was verified by anomaly detection of 1000 registered information in a certain region.