Utilizing Dynamics Patterns of Trust for Recommendation System

Abstract

Our trust always changes due to various conditions. We call it dynamics of trust. Researching about the change of trust is a necessary task because it could be applied into diverse applications such as trust-based recommendation systems. In this paper, our target is to i) propose a general definition of trust, ii) define temporal and spatial patterns for representing how trust changes. They can be used as the output of trust-based recommendation systems, and iii) proving the effectiveness of our proposition on a case study between a Twitter user and a restaurant. Moreover, this research could be applied to various areas (e.g., e-leaning environment and search system).