Generalized Discriminant Analysis algorithm for feature reduction in Cyber Attack Detection System

TL;DR

This paper introduces a novel Generalized Discriminant Analysis (GDA) algorithm for feature reduction in cyber attack detection, improving classification accuracy and efficiency over traditional methods by effectively handling non-linear data.

Contribution

The paper proposes a new GDA-based feature reduction technique specifically designed for cyber attack detection, addressing the limitations of linear methods like LDA.

Findings

GDA outperforms LDA in non-linear datasets.

Reduced feature set improves classifier accuracy.

Decreases training and testing time for classifiers.

Abstract

This Generalized Discriminant Analysis (GDA) has provided an extremely powerful approach to extracting non linear features. The network traffic data provided for the design of intrusion detection system always are large with ineffective information, thus we need to remove the worthless information from the original high dimensional database. To improve the generalization ability, we usually generate a small set of features from the original input variables by feature extraction. The conventional Linear Discriminant Analysis (LDA) feature reduction technique has its limitations. It is not suitable for non linear dataset. Thus we propose an efficient algorithm based on the Generalized Discriminant Analysis (GDA) feature reduction technique which is novel approach used in the area of cyber attack detection. This not only reduces the number of the input features but also increases the classification accuracy and reduces the training and testing time of the classifiers by selecting most discriminating features. We use Artificial Neural Network (ANN) and C4.5 classifiers to compare the performance of the proposed technique. The result indicates the superiority of algorithm.