Mining trust and distrust relationships in social Web applications

• Published in: 2010 IEEE International Conference on Intelligent Computer Communication and Processing pp. 73 - 80
• Main Authors: Zolfaghar, Kiyana, Aghaie, Abdollah
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.08.2010
• Subjects: Communities; Computational modeling; Context; Distrust; Machine learning; Prediction algorithms; Predictive models; Sign prediction; Social network services; Social web; Trust
• ISBN: 9781424482283; 1424482283
• DOI: 10.1109/ICCP.2010.5606460

By the immense growth of social applications in web environment, the role of trust in connecting people is getting more important than ever. Although many researchers have already conducted comprehensive studies on the trust related applications, the understanding of distrust relations is still unclear to the researchers. In this paper, we have investigated some of mechanisms that determine the signs of links in trust networks which consist of both trust and distrust relationships. Achieving this, we develop a framework of trust sign prediction, taking a machine-learning approach. We report experiments conducted on Epinions which is a well-known and very large collection of data dealing with trust computation. Empirical results show that the sign of relations in the trust networks can be effectively predicted using pre-trained classifiers.