Analysis of User Suspicious Behavior in Power Monitoring System Based on Text Vectorization and Unsupervised Machine Learning

In order to solve the feasibility, timeliness and accuracy of the power monitoring system’s user safety behavior analysis, a user suspicious behavior analysis method for power monitoring system was proposed by combining TF-IDF and k-means++ algorithm, and the method was run on Spark, a big data analysis and calculation engine. Experiments show that this method has good performance and can effectively analyze suspicious behaviors of users with potential threats.