Design of a New Efficient Hybrid System for Intrusion Detection Based on HSM Fuzzy Decision Tree

Abstract

Intrusion detection is a primary component of internet security mechanisms. It requires various issues (e.g., accuracy, run-time and memory efficiency) to be improved for analyzing a large amount of network data. This paper proposes a new efficient hybrid system to build an intrusion detection model. By virtue of B. Chandra's heterogeneous node split measure (HSM), we employ principal component analysis, K-means clustering and HSM-based fuzzy decision tree algorithm to construct the system. We discuss approaches as well as the credibility for improving accuracy and efficiency of the detection model. This paper povides the key ideas and discusses the effectiveness of our proposed system.