Intrusion Detection using Feature Selection and Hybridized

Recently, it is adamant to maintain a secured and trusted information transformation between various organizations. Inarguably, many of security technologies (data encryption, intrusion prevention etc) are used to prevent network based systems; many still undiscovered.This paper presents an overview of intrusion detection and a hybrid (naïve baye and KNN) classification algorithm. The data set is passed through the naïve baye for classification, generating the prior and conditional probabilities for each example. If misclassification occur, the example will be passed to the KNN which then ranks the neighbourhood and the resulting are weighted using the similarity of each neighbour of the it, if the Sim(X,Dj) is equals to 1, then X is normal else the algorithm finds the K largest Sim(X,Dj), checks it against a threshold. Experimental results show that the hybrid classifier gives the best result in terms of accuracy and efficiency compared with the individual base classifiers.