State grid office system user clustering analysis based on K-means algorithm
Ting Li, Peng Ye, Shikai Zheng · 2018
Collaborative office system of State Grid have large numbers of user and many application modules, Different types of users have different operation behavior for these modules. The user interest domains based on operation log data uses clustering such as k-means to analysis the users, then for each group of users, analyze it and give it tags. After the analysis, the users can be clustered based on their behavior characteristic. Firstly, the concept of silhouette coefficient is introduced and the optimal clustering number Koptincluded in data set of unknown class information in decided. Then clustering is completed through K-means clustering.