Adaptive Vulnerability Prioritization

From The Math Club

A thing that all vulnerability management companies have failed to currently do is come up with a suitable or adaptive manner of prioritizing and ranking hosts and vulnerabilities after vulnerability assessment scans in a manner that suitably optimizes the work performed by the poor people who have to go and remediate the problems on the network.

None of the proprietary scoring methods provided by any of the assessment companies are very useful. They're all pretty much ad hoc and precariously designed. Even CVSS (Common Vulnerability Scoring System), athough being very adaptive to any situation, requires a very manual and labor intensive process, again relying on the haphazard human to fiddle with numeric scores.

What I am going to present is a few adaptive methods for vulnerability prioritization based on remediation feed back and and a few clustering algorithms based on KNN classification, bayesian analysis and a few other adapative learning methods.