Analysis of Malicious Intruder Threats to Data Integrity

Abstract

The insider threat has increasingly become the cyber security challenge that threatens organisation, financial enterprises, and governmental agencies. Insider threat is being carryout by former and current employees. Meanwhile, insider threat had authorised access to an organisation asset, thus they have better opportunity to undermine the confidentiality, availability, or data integrity than an external attacker. The detection process consists of different techniques, including detecting suspicious activities in the system. This paper focus on insider threat detection through behavior analysis of user’s activities. A deep machine learning approach has been proposed to detect insiders’ threat with better accuracy with low false positive rate. The publicly available dataset used is the CMU CERT synthetic malicious insider threat dataset r4.2. Our empirical evidence outperforms compared to similar existing models, it proved that our approach (LMT) has high accuracy (99.6%), precision (99.6%) and ROC (99.6%).