Hybrid Tabu K Means Clustering for Data Leak Identification

Abstract

Today, there has been an explosion of Online Social Networking (OSN) that has resulted in damages to various organizations owing to data leakage by employees. The social networking behavior of employees, whether intentional or accidental, has given an opportunity for some persistent threats from attackers that realise techniques of social engineering aside from zero-day exploits that are undetectable. Clustering is normally achieved by means of finding any similarity among data for predefined attributes. Similar data are grouped into clusters. In this work, modified methodology of K-Means clustering employed to identify the groups based on common traits. In this work, the Genetic Algorithm (GA) with K-Means, Particle Swarm Optimization (PSO) with K-Means, hybrid GA-Tabu Search (TS) with K-Means and hybrid PSO-TS with K-Means methods are proposed. The proposed hybrid algorithms are used to find better performance for data leakage identification in the social network. The experimental results have demonstrated that the proposed technique performs better in comparison to other methods.