Certain strategic outlier analysis in social networks using graph based approaches

Abstract

In the current big data era, social networking platforms have become the fuel for data proliferation as well as high-impact social influence among people. Due to the simplicity of user account creation and complex user interconnections, social networks have become vulnerable to malicious activity. Malicious users are referred to as anomalies, outliers, bogus or fake users. The outliers percolate through the network community in a disguising manner. Henceforth, the challenging aspect is to find a boundary of separation between fake users and genuine users. The main objective of the research work focuses on determining useful patterns to detect anomalies. Due to the multiple interdependencies and correlations that prevail in social networks, graph representation is preferred. Adopting feature extraction and community detection to achieve scalability and prediction accuracy is the objective of the research. newline