Bots: Genuine or Malicious
Lale Madahali, Jin Tian · 2023
Social network platforms have become an integral facet of contemporary life, serving various purposes such as entertainment, news dissemination, advertising, and personal or corporate branding. Nonetheless, the proliferation of automated accounts, commonly referred to as social bots, poses a considerable challenge, muddying the waters of online reliability. Of particular concern are malicious social bots, which actively manipulate public opinion. In this paper, we introduce the novel concept of ”genuine bots,” which serve constructive roles within the Twitter sphere. We explore the distinctive features of genuine bots, juxtaposing them with their malicious counterparts. Our study culminates in the development of a robust three-way classification model distinguishing between humans, malicious bots, and genuine bots on Twitter. Notably, our experimental findings underscore Random Forest as the most accurate among the various machine learning models assessed. The implications of this work are profound, given the global escalation in bot utilization, promising significant advancements in social network analysis and the broader online sphere.