A cognitive based Intrusion detection system

TL;DR

This paper introduces a new intrusion detection approach combining Deep Neural Networks and SVM classifiers, inspired by divide and conquer, achieving higher accuracy than existing methods on the KDD99 dataset.

Contribution

It proposes a novel hybrid model using deep neural networks and SVMs for intrusion detection, addressing low detection precision and stability issues of previous ANN-based systems.

Findings

Achieved 95.4% classification accuracy on KDD99 dataset.

Outperformed traditional ANN-based intrusion detection methods.

Demonstrated improved detection stability and precision.

Abstract

Intrusion detection is one of the important mechanisms that provide computer networks security. Due to an increase in attacks and growing dependence upon other fields such as medicine, commerce, and engineering, offering services over a network and maintaining network security have become a significant issue. The purpose of Intrusion Detection Systems (IDS) is to develop models which are able to distinguish regular communications from abnormal ones, and take the necessary actions. Among different methods in this field, Artificial Neural Networks (ANNs) have been widely used. However, ANN-based IDS encountered two main problems: low detection precision and weak detection stability. To overcome these problems, this paper proposes a new approach based on Deep Neural Network ans Support vector machine classifier, which inspired by "divide and conquer" philosophy. The proposed model predicts the attacks with better accuracy for intrusion detection rather similar methods. For our empirical study, we were taking advantage of the KDD99 dataset. Our experimental results suggest that the new approach enhance to 95.4 percent classification accuracy.