A Novel Hybrid Feature Selection and Intrusion Detection Based On PCNN and Support Vector Machine

In this paper proposed a hybrid model for feature selection and intrusion detection. Feature selection is important issue in intrusion detection. The selection of feature in attack attribute and normal traffic attribute is challenging task. The selection of known and unknown attack is also faced a problem of classification. PCNN is dynamic network used for the process of feature selection in classification. The dynamic nature of PCNN select attribute on selection of entropy. The attribute entropy is high the feature value of PCNN network is selected and the attribute value is low the PCNN feature selector reduces the value of feature selection. After selection of feature the Gaussian kernel of support vector machine is integrated for classification. Our detection rate is very high in compression of other neural network model such as RBF neural network and SOM network. For the empirical evaluation used KDDCUP99 dataset and measure detection rate precision and recall of proposed model.