Abstract:To solve the problem of low detection performance of the K-means anomaly detection method,an anomaly detection algorithm combining information entropy and improved K-means is proposed.The algorithm uniformly chooses the data object whose density is greater than the average density of the data set as the initial clustering center,avoiding the random selection of the initial center.Besides,the weighted Euclidean distance between the data point and the cluster center in the cluster is calculated according to the attribute weight based on the information entropy.Anomaly detection is performed by comparing the weighted Euclidean distance of the data point with the average weighted Euclidean distance of all data points in the cluster.Experiments show that the improved algorithm has higher detection rate and lower false detection rate.When the algorithm is applied to power load data,the detection rate reaches 90.5%.The abnormal power load data can be effectively detected.