COMPUTATIONAL STRATEGIES FOR TOPIC TRUST PROPAGATION BASED ON K-LEVEL NEIGHBORS

Abstract

Topic trust in social networks is defined by means of a function of trust degrees, which are estimated via interaction experience and user interests. The computation of such a function is based on propagation of trust values along paths with neighbor nodes and thus own highly computational cost. In this paper, we first consider various strategies for estimating topic trust based on a hierarchy of users with k-level neighbors. Then we introduce algorithms for computing topic trust values
w.r.t. these strategies.