Performance Comparison of Intrusion Detection System Classifiers Using Various Feature Reduction Techniques

In this paper, we compare the performance of Intrusion Detection System Classifiers using various feature reduction techniques. To enhance the learning capabilities and reduce the computational intensity of competitive learning neural network classifiers, different dimension reduction techniques have been proposed. These include: Principal Component Analysis, Linear Discriminant Analysis, Independent Component Analysis. Many Intrusion Detection Systems are based on neural networks. However, they are computationally very demanding . In order to mitigate this problem, dimension reduction techniques are applied to a given dataset to extract important features. In the proposed research various classifiers are applied to the reduced feature dataset and their performance is compared. On the basis of these results, a technique is proposed which performs exceptionally well, in terms of both accuracy and computation time. When applied to the KDDCUP99 reduced feature dataset, this technique performs better than a standard learning schema based on the full featured dataset.