Machine Learning Models for Detecting Fake Profiles on Social Media Platforms

Abstract

Social media platforms have become integral to modern communication, yet they face a persistent challenge from fake profiles that undermine user trust and platform integrity. This research investigates the application of machine learning models to identify and classify fraudulent accounts across major social networking sites. The study employs a mixed-methods approach, combining quantitative analysis of profile characteristics with qualitative assessment of behavioral patterns. We developed and tested multiple machine learning algorithms including Random Forest, Support Vector Machines, Neural Networks, and Gradient Boosting techniques on a dataset comprising 15,000 authentic and fabricated profiles. Our findings reveal that ensemble methods achieve detection accuracy rates exceeding 94%, with behavioral features proving more discriminative than static profile attributes. The research identifies key indicators of fake accounts, including irregular posting patterns, suspicious follower-to-following ratios, and anomalous engagement metrics. These insights contribute to the growing body of knowledge on digital security while offering practical solutions for platform administrators seeking to protect their user communities from malicious actors.