User Role Discovery Method in Virtual Community Based on DBSCAN Clustering
Chi Xu, Hua Yong Xiao, Zhongyuan Hua, Xianteng Gu · 2022
With the development of computer network and the wide use of smart phones, people can easily share the knowledge and information they are interested in to each other, and care for each other as if they were friends, thus a virtual community has emerged. How to discover users' social attributes in virtual communities is an important issue of interest to community managers. In this paper, we propose a method to discover user roles from user check-in data in virtual community, the method consist of two parts, the first part vectorizes user feature from user check-in data, and the second part using DBSCAN cluster algorithm to discover user roles from their features of vectorization. An experiment was conducted on a public available data set, and 16 user roles from 3 kind of user feature were discovered, which have varified the effectiveness of the method.