8th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing

Research Article

Robust Expert Ranking in Online Communities - Fighting Sybil Attacks

Download688 downloads
  • @INPROCEEDINGS{10.4108/icst.collaboratecom.2012.250439,
        author={Khaled Ahemd Naji Rashed and Cristina Balasoiu and Ralf Klamma},
        title={Robust Expert Ranking in Online Communities - Fighting Sybil Attacks},
        proceedings={8th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing},
        publisher={IEEE},
        proceedings_a={COLLABORATECOM},
        year={2012},
        month={12},
        keywords={mhits algorithm robust expert ranking collaborative fake detection fighting sybil attacks trust- awarenes},
        doi={10.4108/icst.collaboratecom.2012.250439}
    }
    
  • Khaled Ahemd Naji Rashed
    Cristina Balasoiu
    Ralf Klamma
    Year: 2012
    Robust Expert Ranking in Online Communities - Fighting Sybil Attacks
    COLLABORATECOM
    ICST
    DOI: 10.4108/icst.collaboratecom.2012.250439
Khaled Ahemd Naji Rashed1,*, Cristina Balasoiu1, Ralf Klamma1
  • 1: RWTH Aachen University
*Contact email: rashed@dbis.rwth-aachen.de

Abstract

Nowadays, many online communities provide means for users to contribute in the evaluation of community created media by tagging, commenting and rating. Judging the users expertise in such collaborative systems is an important issue. As these systems are becoming increasingly popular, they are attackable, e.g. by Sybil Attacks. Thus, an effective expert ranking strategy must be robust to such attacks. In this paper, we propose MHITS, an algorithm to rank users’ expertise by exploiting the number of users’ fair ratings and direct trust users gain in the online community. We integrate SumUp, a Sybil-resilient algorithm, into MHITS algorithm as a robust ranking strategy. Experimental results show the effectiveness of the proposed method, which can ensure that the highly ranked experts are highly trusted users and provide the high number of fair ratings for the relevant media. We contribute to the experimental evaluation of algorithms for online systems, fighting malicious behavior.