IMPROVING SOCIAL MEDIA PLATFORM INTEGRITY USING MULTI FEATURE-BASED MALICIOUS PROFILE DETECTION

Abstract

The purpose of this work is to enhance the integrity of social media platforms by utilizing a multi-feature-based strategy for detecting harmful profiles. This will allow for the identification and mitigation of coordinated inauthentic accounts, spam accounts, and fraudulent accounts. Because of the proliferation of online interactions, malicious actors can easily take use of social media platforms to spread false information, conduct phishing attacks, influence public opinion, and compromise users' privacy. Behavioral dynamics, account age, username patterns, bio characteristics, and content-based features like posting frequency, sentiment patterns, and hashtag usage are all part of the proposed method. Network-based metrics and interaction graphs are also part of it. Behavioral features like clustering behavior and automation indicators are also part of it. The detection accuracy is improved and the number of false positives is decreased by using advanced machine learning and ensemble classification algorithms. Using a thorough multi-dimensional feature space, this research aims to enhance social media ecosystems' trust, transparency, and security in order to foster safer online communities and more reliable digital communication environments.