Perceptual intrusion detection system

Abstract

A simple intrusion detection system (IDS) with respect to perception of human experts is proposed. Its computational framework is designed based on concepts of computational theory of perceptions (CTP) and mass assignment theory (MAT). CTP provides a computational framework of representing and handling perception using linguistic terms and corresponding fuzzy sets. MAT provides that of representing and handling a bias consisting in between perception and observed intrusions through the consistency management between fuzzy sets and probability distributions. For the knowledge construction of this IDS, only linguistic descriptions extracted from organization policies and perception of human experts is expected. For the inference and refinement of knowledge, support logic, a truth functional logic for manipulating probability intervals, is used.