The Effect of the Similarity Between the Two Names of Twitter Users on the Credibility of Their Publications

摘要

Social media networks have become one of the most important sources for information and news sharing. As a result of the ease of publication and the lack of censorship of the published content, they have become a source of spreading rumors and deceptive information. Twitter, as a social media site, has its importance in influencing the world public opinion in general and the Arab street in particular. Many studies have been conducted to examine the credibility of the publication through social networking sites, especially through Twitter in different languages, but the research in Arabic is few and needs further development. In this research, we focus on discovering the credibility of Twitter publishing in Arabic by integrating the Social Media Mining techniques with the Natural Language Processing (NLP). We have extracted the common 40 user and content features for training and testing the most used four classifiers in field of credibility detection. In order to enhance the accuracy of the detection process, we developed a new algorithm to discover the relationship (similarity) between the two names of Twitter user (username and display name). Using the random forest classifier, we outperform the baseline works in terms of accuracy.