Scam Shield

Abstract

This project report explores the progression of machine learning (ML) applications in the detection of fake accounts on social media platforms. While ML's roots in this domain trace back to the late 2000s, its widespread adoption for this purpose gained momentum in the 2010s. The early breakthrough, exemplified by a 2009 study at Carnegie Mellon University, demonstrated the superior accuracy of ML classifiers in distinguishing fake accounts on MySpace compared to conventional methods. Subsequent years witnessed continuous refinement of ML algorithms for fake account detection, incorporating techniques such as classification, anomaly detection, and graph-based algorithms. Major social media platforms, including Facebook and Twitter, embraced ML. to identify and remove fake accounts, marking a significant improvement in detection capabilities. The adoption of ML. resulted in the removal of millions of fake accounts annually, a substantial advancement from previous methods