Trust and reputation model for various online communities

Abstract

In World Wide Web there are many online communities with a huge number of users and a great amount of data which are continuously increasing. In this context it is important for users to interact with resources and other users according to their preferences. On this direction of information filtering domain our work is oriented to trust based filtering. We have developed a model formed by three interconnected components: trust component which allows computation of trust levels among users which are not directly connected, a component which computes reputation of entities in the system, and a recommendation component. Several sets of tests of our model have been performed and we have the possibility to integrate it in various online communities.