A statistical approach for reducing misinformation propagation on twitter social media

Abstract

Misinformation on Twitter has caused havoc all over the world, thus posing a prime cybersecurity concern. Hence, our research objective is to tackle the menace of infodemic by computing the credibility of Twitter users. Our proposed approach rates the Twitter users on a scale of credibility rating, obtained by analysing the pattern of growth rate statistics of the tweeting and following behaviour. Hence, our research provides two types of credibility computational frameworks that includes Twitter-user profile-centric activities and tweet-propagation features. Initially, we establish and justify the fact that there is a observable difference in the propagation pattern of rumored and non-rumored tweets by analysing the posts of 129,323 Twitter users. Later, the credibility score is devised based on tweet propagation pattern. It contains parameters that satisfies a statistical significance level of 99.99%. The second credibility score formula utilizes more profile centric features and almost all the parameters used for second formula satisfy a statistical significance level of 95%. Our research will significantly assist social media administrators to identify the rumor spreaders, thereby reducing the menace of misinformation.