X (Twitter) Bot Detection Utilizing Machine Learning

Authors: Amanda Harnos, Nur Dean, Kathleen Tan

SUNY Campus: Farmingdale State College

Presentation Type: Poster

Location: UU 108

Presentation #: 90

Timeslot: Session C 1:45-2:45 PM

Abstract: In today's digital age, social media has become an integral platform for disseminating information while raising cybersecurity concerns. However, alongside its widespread popularity, there has been a notable rise in the prevalence of botnet and bot threats. The bots controlled by a botnet are more likely to have malicious intent and an overall threat to social media users' cybersecurity with populated misinformation and security threats. The network of bots is capable of spreading phishing links to fraudulent websites and manipulating information for political or social gain. Of particular interest is X (Twitter), a social media platform known to harbor a plentiful amount of bot accounts that threaten the integrity of the platform. In light of this, our study aims to explore methods for identifying these bots and assess the impact of various features in aiding their detection. The data derived was carefully curated and filtered through standardization and regularization mathematical procedures. Within the research completed, various machine learning methods for supervised machine learning were employed, including linear regression, logistic regression, support vector model, and k-fold cross-validation. The significance of social bot detection should not be an issue to ignore. While the importance continuously increases along with the threat to social media users' security.