FP-ANK: An Improvised Intrusion Detection System with Hybridization of Neural Network and K-Means Clustering over Feature Selection by PCA

Intrusion Detection System (IDS) predominantly works for detecting malicious attacks. Many researchers have proposed the IDS with different techniques to achieve the best accuracy with the consolidation of Clustering and Artificial Neural Network (ANN). Clustering and ANN based models give better precision rate with better accuracy where attack records are low. Nevertheless, all the features of dataset are not relevant for classifying different attacks. So, feature selection can improve the stability and accuracy of IDS. In this paper, it is proposed that IDS with the amalgamation of best efficient features selected by Principal Component Analysis (PCA) can reduce the computational complexity of the system. It has been combined with the K-means clustering technique to cluster the specific groups of attacks and Artificial Neural Network to get a preeminent output by training the formulation of different base models. The model name has been defined by FP-ANK model. Investigational results have been reported on the NSL-KDD dataset where the accuracy rate associating with other models is distinct to validate the proposed system.