Collaborative Filtering Based Spammer Detection and User Reputation Estimation in Rating System

DOI

10.2991/etmhs-15.2015.23How to use a DOI?

Keywords

Rating system; User reputation; Collaborative filtering; Spammer detection

Abstract

Online rating systems play vital role in the recommendation systems and deeply influence the following user choice. It’s common occurrence that spammers contaminate the rating systems with malicious rates. However, most of the researchers pay more attention to the accuracy and capability of user preference prediction, rather than the authenticity to rating systems, which provide basis and resource to the recommendation systems. Taking advantage of recommendation algorithm research achievement, we propose a collaborative filtering based spammer detection and user reputation estimation method which perfectly resolves the sparsity problem in huge rating data and promotes poor user preference property of item-based user reputation algorithms. The experiment shows effectiveness of the algorithm both in traditional mode of spammer attack and newly proposed one in this paper, which highly simulates the behavior of real word spammers.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Chaojun Liu
AU  - Junyu Niu
AU  - Fangjun Liu
PY  - 2015/03
DA  - 2015/03
TI  - Collaborative Filtering Based Spammer Detection and User Reputation Estimation in Rating System
BT  - Proceedings of the 2015 International Conference on Education Technology, Management and Humanities Science
PB  - Atlantis Press
SP  - 100
EP  - 104
SN  - 2352-5398
UR  - https://doi.org/10.2991/etmhs-15.2015.23
DO  - 10.2991/etmhs-15.2015.23
ID  - Liu2015/03
ER  -