Prioritizing Vulnerabilities with Context-Based Detection Model Using Machine Learning in Healthcare Industry
Open Access DepositedDownloadable Content
In the constantly changing healthcare landscape, the security of digital healthcare systems is paramount, given healthcare data's sensitivity and critical nature. This Praxis introduces a practical approach to enhance the cybersecurity posture of healthcare industries by prioritizing vulnerabilities through context-based detection while utilizing machine learning techniques. This Praxis proposes Exploit Score (ETS), a solution that identifies and ranks vulnerabilities based on their severity, common vulnerability exploit (CVE), access type, device type, device function, open ports, ransomware campaign, and exploitability with the integration of machine learning algorithm for timely vulnerability remediation and remediation of healthcare focused critical vulnerabilities. The implemented approach improved the current CVSS scoring system by focusing on healthcare's most critical systems and their timely vulnerability remediation, leading to the on-time remediation of most critical healthcare devices. The solution employs a context-based detection framework called Exploit Score (ETS), which leverages machine learning algorithms to analyze patterns and behaviors within detected vulnerabilities and identify deviations that could be used to prioritize vulnerabilities based on their importance to the healthcare organization. Using a dataset that employs real-world vulnerability scan results, we treated the dataset by removing features that do not apply to the requirement. We further added context-based features to complete the requirement. We initially trained and tested different models to gather further information and then evaluated their effectiveness in prioritizing vulnerabilities. Then, we arrived at the top two performing models: the XGBoost and CatBoost regression. Our result significantly improved over previously established vulnerability prioritization and remediation research work. The research concludes that ETS context-based integration with machine learning can be deployed in healthcare organizations to help prioritize vulnerability remediation promptly while focusing on healthcare-specific critical vulnerabilities and ensuring healthcare critical vulnerabilities are remediated first. Among the four machine learning models investigated, eXtreme Gradient Boosting shows the best performance with the highest R2 score, lowest MAE, and the fastest training time.
- 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.