A model for recalibrating credibility in different contexts and languages--a Twitter case study

Abstract

<p>Due to the growing dependence on the WWW UserGenerated Content (UGC) as a primary source for information and news, the research on web credibility is becoming more important than ever before. In this paper we review previous efforts to evaluate information credibility, focusing specifically on micro-blogging. In particular, we provide a comparison of different systems for automatic assessment of information credibility based on the used techniques and features, and we classify the Twitter credibility surveys based on the features considered. We then propose a general model to assess information credibility on UGC different platforms, including Twitter, which employs a contextual credibility approach that examines the effect of culture, situation, topic variations, and languages on assessing credibility, using Arabic context as an example. We identify several factors that users may consider in determining credibility, and argue that the importance of each factor may vary with a context. Future work will include both a user study and machine learning techniques to evaluate the effectiveness of various factors for information credibility classification in different contexts.</p>