Enhancing Social Network Security: Machine Learning-Based Bot Detection

Abstract

The proliferation of social media platforms has transformed communication, but it has also given rise to social media bots that can spread misinformation, manipulate public opinion, and compromise the integrity of online discourse. This thesis addresses the critical issue of detecting social media bots on Twitter. Traditional detection methods often fall short due to the evolving nature of these bots and the vast amount of data involved. To overcome these challenges, this research proposes a hybrid ensemble model that combines profile-based and content-based features with advanced natural language processing techniques. This approach captures a wide range of bot behaviors and characteristics, resulting in more accurate and robust detection. The thesis includes an examination of social media platforms and the threats posed by bots, a review of current bot detection methods, an in-depth explanation of the proposed hybrid ensemble methodology, and an experimental evaluation of the methodology’s effectiveness compared to leading techniques. The findings demonstrate significant improvements in detection performance, supporting efforts to protect social media environments from harmful automated entities.