Hamda Slimi

Affiliation
Position: 
PhD Student
Institution: 
Higher School of Computer Sciences (ENSI)
University: 
La Manouba University
Research laboratory: 
Research
Research Profile: 
PhD
Supervisors (s)
Co-upervisor(s)

Ph.D Thesis description

Brief description

Information on social media platforms suffers from a relative absence of professional gatekeepers to monitor content. Thus evaluating information credibility on these platforms has become a vital issue for today information consumers. The main focus of our research is the credibility of information spread through social media networks, we aim at developing a new approach for evaluating the credibility of tweets. In this approach, we introduce new features to the evaluation of credibility on Twitter, by assigning a level of credibility to the topics discussed on Twitter and calculating the expertise of a user in a certain topic. 

Detailed description

Social media can provide instant news faster than traditional news outlets or sources and can be a great source of information, but there is an increasing need to verify and determine the accuracy and truthiness of this information. The credibility of information on social media depends mainly on three criteria.

First, source credibility which depends on the expertise and the likelihood of the source to provide credible information. Secondly, content credibility or message credibility which refers to the informational quality and accuracy. Thirdly, the credibility of the medium on which the information is communicated.

Since credibility depends tremendously on the context on which it will be studied, we narrowed our research to a particular social media platform. Twitter offers a multitude of features that can be exploited to determine the credibility of information disseminated through the network. In our research, we aim at enhancing the accuracy of credibility prediction by introducing a novel approach that takes advantage of the features offered by Twitter and combines them to obtain a new perception on credibility criteria.