Abstract
Twitter is an important platform for many activities such as communication, promotion and news. Twitter is starting to clamp down on removing content and accounts whose sole purpose is to spread disinformation at scale. The increasing number of bots on Twitter has led to a significant circulation of malicious content. It has the power of manipulating and influencing people negatively. Hence bot detection and finding their extent of influence is of prime importance. This research has used varied approaches for the detection of bots; a supervised machine learning approach which makes use of a unique feature called the bot score to determine the bot probability of a user. Secondly, using an unsupervised machine learning approach, users were divided into various clusters based on their activity. Furthermore, to assess the influence of users in hashtag manipulation, this research analyzed various categories of users in a hashtag and made promising conclusions.
