A trust prediction framework in rating-based experience sharing social networks without a Web of Trust

As online experience sharing sites have become one of the popular collaborative online communities, people are easily able to share their good and bad experiences on various products and services with a large number of unknown people as well as their friends. These experience sharing communities try to encourage social interaction among people and facilitate experience sharing and dissemination with satisfaction. The social interactions among users in such online communities are constructed based on trust that is established from each user's subjective perspective on the experiences in the community. Since a robust trust system is vital in experience sharing online communities, we therefore propose a computational trust framework for predicting a degree of trust or trust-connection between a pair of users. The Web of Trust, which consists of explicit trust rating among users, is not always available and is typically sparse, so the proposed framework does not rely on a Web of Trust. The proposed trust system measures a degree of trust based on users' expertise and preferences regarding topics (i.e. categories), using users feedback rating data which are available and much denser than a Web of Trust. In order to derive a more personalized degree of trust, the expertise- and preference-based trust is refined with each user's subjective and direct experiences with community members as well as a target user. The empirical experiments show that our proposed trust framework is quite promising in ratings-based online experience sharing communities, even when there are not enough user feedback ratings to predict a degree of trust.