Detection of Malicious Bots in Twitter Using Machine Learning Algorithms

Abstract

Unwanted social bots have become more pervasive as robotized social entertainers because of the development of web administrations and the ubiquity of online informal organizations (OSN) like Facebook, Twitter, LinkedIn, and so on. These players can assume a variety of malevolent roles, such as intruders into human discussions, con artists, imposters, spreaders of false information, manipulators of the stock market, astroturfers, and any content polluters (spammers, virus spreaders, etc.). Social bots are undeniably quite important on social networks. The RNN algorithm feeds the output of the previous step as input to the current step in order to detect these harmful social bots in social networks. Comparing this algorithm to all other machine learning algorithms, it has a high detection accuracy rate. Therefore, this study examines detection strategies within a methodological category, highlights the possible risks of malevolent social bots, and suggests lines of future research.