FAKE PROFILES IDENTIFICATION IN ONLINE SOCIAL NETWORKS USING MACHINE LEARNING AND NLP

Abstract

Online Social Networks (OSNs) have become an integral part of modern communication, connecting millions of users worldwide. However, the proliferation of fake profiles poses significant security and privacy threats, including misinformation, fraud, and cyber harassment. Detecting and mitigating these fraudulent accounts is essential to maintaining the integrity of online platforms. Traditional rule-based approaches struggle to adapt to the evolving tactics of malicious actors, necessitating more robust and intelligent solutions. Machine Learning (ML) and Natural Language Processing (NLP) offer promising avenues for accurately identifying fake profiles by analysing behavioural and textual features. This study explores the application of ML and NLP techniques for the detection of fake profiles in OSNs. A comprehensive dataset of both genuine and fake accounts is utilized to extract key features such as profile attributes, activity patterns, and linguistic characteristics of user generated content. Various supervised and unsupervised ML models, including decision trees, support vector machines, deep learning networks, and ensemble methods, are evaluated for their effectiveness in identifying fraudulent accounts. Additionally, NLP-based techniques, such as sentiment analysis and text embedding models, are leveraged to analyse user generated content for signs of deception and unnatural linguistic patterns.