Simulated Dynamic Access Control (Simulated DAC) Evaluation with OCEAN-Based Behavioral Profiling for Insider Threat Detection in U.S. Healthcare
Open Access DepositedDownloadable Content
Insider threats are becoming a significant problem for U.S. healthcare systems, astraditional static access controls often fail to detect subtle, evolving user behaviors. This praxis presents a Simulated Dynamic Access Control (Simulated DAC) framework that combines behavioral analytics with an OCEAN-based psychometric profile to facilitate the identification of threats. The evaluation was based on three research questions. RQ1 examined whether the Simulated DAC framework reduces the number of false negatives. RQ2 examined how the addition of psychometric traits affected the accuracy of detection. Forest of Random Stratified cross-validation, SMOTE, and temporal splits were used to train the RF and XGBoost (XGB) models. Finally, RQ3 used role deviation scoring to see how well the model could generalize over time. The Simulated DAC models significantly reduced false negatives (up to 54.35% for XGB), while also improving recall (≥ 0.53) and precision (≥ 0.83). The ROC AUC values were 0.9895 for RF and 0.9889 for XGB. The framework got 87.91% accuracy on future data for RQ3, which is statistically better than the 85% hypothesis limit. Statistical tests revealed strong benefits (p < 0.001), and SHAP analysis confirmed that the model was clear. These results demonstrate that Simulated DAC is a scalable, flexible, and user-friendly approach for identifying insider threats in evolving healthcare settings.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.