Evaluation of Credibility Assessment for Microblogging – Models and Future Directions

Abstract

<p>Due to their openness and low publishing barrier nature, UserGenerated Content (UGC) platforms facilitate the creation of huge amounts of inaccurate content. Consequently, assessing UGC information credibility is developing into a vitally important research topic. This paper offers a systematic review of the current developments in assessing information credibility automatically in UGC platforms, focusing on microblogging service. It covers different aspects from dataset collection and feature usage, through classification techniques, to performance evaluation. A novel theoretical credibility model which integrates the evaluators' traits and context factors to assess information credibility is also presented along with important directions for future research on UGC information credibility.</p>