Insider Threat Detection Through Individual to Role and Psychometric Group Distance Features with Supervised Machine Learning Methods

Insider threats are a continually growing concern for organizations that can result in significant negative impacts including financial and reputational damage. Detecting these threats effectively and efficiently is a challenge for organizations given the complexity of the data associated with ecosystems in which the insider operates. Social, economic, and political motivations of various types of malicious actors result in patterns of behavior that are difficult to detect earlier to minimize impact.This praxis proposes a novel detection model that can improve both detection performance and timeliness. Thematic activity vectors are created using the CERT 4.2 dataset given by Carnegie Mellon University to determine the distance to groups in which a user participates. Others in the same role have been researched previously; however, the construction of distance vector features can also prove to be useful for detection performance. Moreover, the creation of psychometric groupings can also be used to calculate vector distances from individual user activity. With these new features, augmentation to user rolling behavioral activity which captures several statistical viewpoints will show improvement. Supervised machine learning methods will be leveraged to explore the possible benefits of inclusion of this multiple vector distance measurement approach.

Author

• Clements, Kenneth

Language

• English

Keyword

• Psychometric data
• Insider Threat
• Machine Learning

Date created

• 2024

Type of Work

• Dissertation

License

• All rights reserved

Rights statement

• In Copyright

GW Unit

• Cybersecurity Analytics

Degree

• D.Engr.

Advisor

• Hashemipour, Medhi

Committee Member(s)

• Hashemipour, Medhi
• Fossaceca, John
• Hassan, Sayed

Persistent URL

•  https://scholarspace.library.gwu.edu/etd/hh63sw71w