Machine Learning Based Automated Detection of Fake Twitter Accounts

Abstract

Online social networks (OSNs) are becoming increasingly popular as communication tools to enable a wide range of individuals and organisations to publish and share news and information. However, the ubiquity of OSNs also means they have become a key target for spammers to disseminate malicious content. Twitter, the micro-blogging platform, is among the largest OSN in use nowadays. It enables users to share and seek a wide variety of current information. Thus, Twitter’s large user base is attractive for spammers to use fake accounts to spread malicious content such as phishing scams, fake advertisements, and fake news. To address this issue, machine-learning (ML) algorithms can be used to detect Twitter accounts actively engaged in spreading spam content. Therefore, the present research project aims to use three ML algorithms, namely, K- nearest neighbour, random forest, and support vector machine, to experiment on a dataset of Twitter accounts to identify such fake accounts.