基于Spark和K-means++的电力数据异常检测OA

With the gradual improvement of power grid information level,power data shows a trend of quantification and com-plexity.Aiming at the problems of low accuracy and poor stability of traditional power big data anomaly detection methods,insuffi-cient computing resources in the context of massive data,and low efficiency of MapReduce engine in Hadoop architecture in the con-text of repeated iterations such as machine learning,a parallel K-means++algorithm based on Spark is proposed.The simulation re-sults show that the algorithm has a good effect on the detection of abnormal data in the power data scenario,and can effectively im-prove the accuracy and convergence of clustering results.