Evaluating quality of Web2.0 UGC based on user authority and topic distribution

User Generated Content (UGC) could be composed and published by every user. As a result, UGC quality may not be well-guaranteed. We put forward a synthetic evaluation method for UGC quality of content via user authority and topic distribution. We consider various features of users, including basic registration information, relative static online social relations and dynamic interactions between users. We adopt Link Analysis (LA) based on user credibility and Generalized Regression Neural Network (GRNN) to rate user authority and then analyze the probabilistic topic distribution of authors and UGC based on textual content using Author-Topic Model (AT). We´ve also integrated the author´s actual behavior during the composing of UGC to compute the contribution degree of user authority. Finally, we combined user authority and contribution degree to evaluate overall quality of a multi-author UGC. Experiments with real dataset collected from TianYa have shown that the proposed method could evaluate UGC reasonably.