A hybrid approach of trust inference and social network analysis for expert search

Abstract

This study combines a trust inference (The ‘Tidal Trust’ algorithm) and social network analysis (SNA) to provide a new way of searching the experts (e.g., physicians) from the blog's articles. The hybrid approach to measure the trust values and centralities of social networking through the interactions of blogs' trackbacks can facilitate searching the trustable blog's articles and expert recommendations for the query users/bloggers. Furthermore, knowledge dissemination and acquirement can also be improved by the bloggers' interactions and the appraisals assigned by the bloggers on the blog's web sites. In order to verify the effect of hybrid approach, the experiments test the variation of recommendation values inferred by trust value and rating based on 50 blogs' articles and 20 blogs' authors. The results demonstrate the high trust value the high recommendation value. In terms of the variation of recommendation values, the high trust value is lower than the low trust value. The quality of recommendation results would be stable. The experimental results demonstrate the rating is also important to influence the recommendation value in this trust-based recommendation methodology.