Power monitoring system log analysis method based on K-Means clustering. Issue 1 (1st October 2022)

• Record Type: Journal Article
• Authors: Pang, Xiaojian; Wang, Binquan; Zeng, Chuyang; Tao, Wenwei; Su, Yang; Hu, Zhaohui
• Is Part Of: Journal of physics. Volume 2354:Issue 1(2022)
• Publication Date: 2022-10-01
• Subjects: Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/2354/1/012008
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: The amount of log data of the power monitoring system is very large, and log analysis is difficult, which causes the power monitoring system to take more time to analyze the log, and there are problems such as low recall rate. Based on this, the design of the power monitoring system log analysis based on K-Means clustering method. For log mining and preprocessing, the K-Means clustering algorithm is used to measure the similarity and dissimilarity of the logs, clustering and locating the dominant number of similar logs in all log data, and compiling standardized documents for them, and compiling the After the data is stored, the log template is established to analyze the power monitoring system log, so as to realize the power monitoring system log analysis based on K-Means clustering. The experimental results show that the log analysis method studied effectively improves the analysis efficiency and recall rate, and can accurately analyze the log, find abnormal logs, reduce the number of false alarms, and prove the effectiveness of the log analysis method.