Rumor Detection In Social Media

Computer Engineering

Social Media has outpaced the conventional news media systems. It is often observed that news stories are first broken in cyber space and then the electronic and print media take them up. However, the distributed architecture and lack of moderation in most of the social media platforms, with the temptation of the users for posting a newsworthy story early on, makes the veracity of information a major issue. This information is defined as rumour, which is a non-credible piece of data circulating in cyberspace, often causing social unrest. For establishing credibility of this information, we define a two phase approach, considering Twitter as the target social networking platform specifically of Indian domain. First phase is based on the premise that verified News Channel Handles in Twitter would furnish more credible information as compared to the nave general public at large. Live streaming or recent tweets are extracted corresponding to Twitter trends, based on clustering using Hashtags. Contextual and sentiment mismatch ratio between the tweets of above mentioned classes is done using semantic and sentiment analysis of tweets,which would reflect the degree of discrepancy of the information. However, as a tweet is restricted to 140 characters, it may not be sufficient to gain useful insight. Tweets are also is susceptible to noise, which may decrease the accuracy of analysis. To overcome these problems, the second phase verifies the claim credibility result, from heterogeneous data resources like authentic web articles, web pages, blog posts, websites, etc. using web crawling techniques. The performance of the proposed method is evaluated, considering the recent examples of popular rumours that went viral in social media.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.