Ranking online quality and reputation via the user activity

作     者：Liu, Xiao-Lu Guo, Qiang Hou, Lei Cheng, Can Liu, Jian-Guo 

作者机构：Univ Shanghai Sci & Technol Res Ctr Complex Syst Sci Shanghai 200093 Peoples R China 

出 版 物：《PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS》 (物理学A辑：统计力学及其应用)

年 卷 期：2015年第436卷

页      面：629-636页

核心收录：

学科分类：07[理学] 0702[理学-物理学] 

基　　金：National Natural Science Foundation of China [71171136, 71371125, 61374177] Shanghai Leading Academic Discipline Project of China [XTKX2012] Shanghai Municipal Natural Science Foundation [14ZR1427800] Foundation of Shanghai Research Institute of Publishing and Media [SAYB1407] Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning 

主　　题：Online rating system User reputation Object quality Iterative algorithm 

摘      要：How to design an accurate algorithm for ranking the object quality and user reputation is of importance for online rating systems. In this paper we present an improved iterative algorithm for online ranking object quality and user reputation in terms of the user degree (IRUA), where the user s reputation is measured by his/her rating vector, the corresponding objects quality vector and the user degree. The experimental results for the empirical networks show that the AUC values of the IRUA algorithm can reach 0.9065 and 0.8705 in Movielens and Netflix data sets, respectively, which is better than the results generated by the traditional iterative ranking methods. Meanwhile, the results for the synthetic networks indicate that user degree should be considered in real rating systems due to users rating behaviors. Moreover, we find that enhancing or reducing the influences of the large-degree users could produce more accurate reputation ranking lists. (C) 2015 Elsevier B.V. All rights reserved.