%0 Journal Article
%A HAN Jing
%A LI Yi-wen
%A LIU Jian-wei
%A Lü Zhi-heng
%A SHUANG Kai
%T Log Template Extraction Algorithm Based on Normalized Feature Discrimination
%D 2020
%R 10.13190/j.jbupt.2019-033
%J Journal of Beijing University of Posts and Telecommunications
%P 68-73
%V 43
%N 1
%X A log template extraction algorithm based on normalized feature discrimination is proposed, aiming at the problem that the number of clusters needs to be provided as a priori information in traditional log template extraction. First, log data is initially compressed to reduce data redundancy. Then, a log clustering process is implemented, and the normalized feature is used to discriminate whether the clustering result meets requirement:if so, the clustering process is successfully completed; if not, the number of log clusters is adjusted by using binary search and redo clustering. Finally, the log template is extracted via clustering results. In addition, an evaluation metric that measures the effectiveness of template extraction is designed. Experiments on real data indicated that the algorithm can achieve more stable and accurate template extraction performance than the benchmark method, and it had good generalization performance.
%U https://journal.bupt.edu.cn/EN/10.13190/j.jbupt.2019-033