Abstract
The use of social media for business and electoral campaigns has increased the amount of harmful behaviour. Social media engagement, like personal, corporate, and political propaganda, has promoted anti-social activity. The intruder receives a full platform to propagate criticism after hiding behind a hacked profile. This research tries to extract relevant elements from social media data indicating a complex connection in order to detect abnormalities. The system successfully identifies abnormal behaviour by recognising malicious or phoney social network accounts. The functional goal of this study is to present a notion of machine learning with a graphical view of social media to detect end-user characteristics. Identify malevolent individuals, their communal information, and sockpuppet nodes, and dramatically enhance classification performance using the end-graphical user's and linguistic viewpoint.