/ml-ac

Machine Learning for Access Control refinement

Primary LanguagePythonMIT LicenseMIT

ML-AC: Adaptive Access Control

This repository contains the algorithm implementation of the ML-AC system presented in this paper.

The solution aims at making access control adaptive by refining at run-time access granted according to monitored behaviours. The solution is building the user profile with a Random Forest algorithm and using a clustering approach to prevent anomaly behaviours to be used in the learning.

Structure of the repository

Contextual Behaviour contains the management of the Random Forest algorithm

Concept Drift contains the clustering approach for handling concept drift

BBNAC contains the Matlab implementation of the approach we used in the experimental validation of ML-AC