Poly Logarithmic Naive Bayes Intrusion Detection System Using Linear Stable PCA Feature Extraction

Software defined network is smart and centralized architecture which increases the network performance and it is efficiently programmed to support different framework of big data and cloud computing virtualization. Several network categorical traffic attacks attributes are many issues which numerous of conventional IDS-Intrusion Detection System with lesser efficiency in terms of recognition, augmented rate of false positive, and bad generalization capacity. Thus, it is necessary to propose a method redresses all of the mentioned issues. In this paper, we propose the IDS methodology to recognize the maliciousness in the Software defined network (SDN) with the novel linearly stable PCA to extract the features. Afterwards, the extracted features will be classified with the novel poly logarithmic function based Naive Bayes classification methodology to diagnose between the normal and abnormal nodes. Finally, we carry out the performance evaluation in terms of accuracy, recall, FPR, TPR, and many performance by using the datasets of \(KDD TEST^{ - 21}\) and KDD TEST plus for validating the proposed IDS performance.