URDO&RG: A Tool for User Role Discovery and Optimization & Requirement Generation Based on User Check-in Data in LBSN

Abstract

The regularities shared by a set of users in a long-term in Location-Based Social Network (LBSN) can be abstracted as user roles. Although methods are proposed to identify user roles, the quantitative analysis of the effectiveness of the method is insufficient, and as far as we know, tools to generate user roles automatically from user data are missing. Therefore, a tool named URDO&RG (User Role Discovery and Optimization & Requirements Generation) is realized. The tool includes four main functions: (1) visualization of the data from multiple dimensions in diverse display diagram, (2) discovery of user roles using various algorithms, (3) optimization of the method to achieve better user role discovery outcome. (4) generation of user requirements from user data. Data set is used to validate the effectiveness of the tool. Result shows that, the tool can discover user roles and generate user requirements effectively, and the optimization method outperforms the original methods in many evaluation metrics.