Machine Learning-Based User Behavior Analytics: Effectiveness in Detecting Account Compromise
APA Citation
Spencer, H. & Yilmaz, E. (2023). Machine Learning-Based User Behavior Analytics: Effectiveness in Detecting Account Compromise. *Computers & Security*. https://doi.org/10.1016/j.cose.2023.103534
View original paper →What Did This Cybersecurity Research Find?
This cybersecurity detection study evaluated ML-based user behavior analytics (UEBA) systems across 6 enterprise environments with injected compromise scenarios. Cybersecurity UEBA systems detected 78% of simulated account compromises within 4 hours, but required a 30-day baseline period and produced elevated false positives during organizational changes like role transitions.
Key Findings
- 1UEBA systems detected 78% of simulated account compromises within 4 hours
- 2A 30-day baseline period was required for accurate anomaly detection
- 3False positive rates spiked during organizational changes (role transitions, mergers)
- 4Combining UEBA with rule-based detection produced the best overall detection rates at 89%
- 5Insider threat scenarios were detected at lower rates (62%) than external compromise (84%)
How Does This Apply to Cybersecurity Careers?
Security analysts working with UEBA tools can understand their capabilities and limitations. Security engineers can set realistic deployment expectations for behavior analytics products.
Who Should Read This?
Frequently Asked Questions
What did this cybersecurity research find?
This cybersecurity detection study evaluated ML-based user behavior analytics (UEBA) systems across 6 enterprise environments with injected compromise scenarios. Cybersecurity UEBA systems detected 78% of simulated account compromises within 4 hours, but required a 30-day baseline period and produced elevated false positives during organizational changes like role transitions.
How is this research relevant to cybersecurity careers?
Security analysts working with UEBA tools can understand their capabilities and limitations. Security engineers can set realistic deployment expectations for behavior analytics products.
Where was this cybersecurity research published?
This study was published in Computers & Security in 2023. The DOI is 10.1016/j.cose.2023.103534. Access the original paper through the publisher link above.
Explore Related Cybersecurity Resources
Get cybersecurity career insights delivered weekly
Join cybersecurity professionals receiving weekly intelligence on threats, job market trends, salary data, and career growth strategies.
Get Cybersecurity Career Intelligence
Weekly insights on threats, job trends, and career growth.
Unsubscribe anytime. More options