Dimension Reduction Features based on Principal Component Analysis in Intrusion Detection Systems

In this paper a new hybrid algorithm PCA-MLP (principal component analysis - Multilayer Perceptron Neural Network) proposed. The PCA transform used to reduce the feature and trained neural network is used to identify the any kinds of new attacks. Test and comparison are done on NSL-KDD dataset. the experiments on NSL-KDD data demonstrate that our proposed with a 36% reduction feature space and the detection rate of 87.65%, better than the same algorithm.

کلیدواژه ها:

Principal component analysis ، dimensionality reduction ، neural network ، NSL-KDD