Modified Naive Bayes Intrusion Detection System (MNBIDS)
Karuna S. Bhosale, Maria Nenova, Georgi Iliev · 2018 International Conference on Computational Techniques, Electronics and Mechanical Systems (CTEMS) · 2018
In this result paper we presenting Modified Naïve Bayes Intrusion Detection System (MNBIDS), it is based on the existing Naïve Bayes system. In that we perform the data pre-processing, data normalization and feature extraction. Now a day's Network security is turning into an expanding vital issue, since the fast advancement of the Web. Information mining and machine learning innovation have been broadly connected to network interruption recognition and anticipation frameworks by finding client behavior standards from the network traffic information. In this system, we use real time packet, which is used to real time analysis and also the KDD Cup 99 dataset for the execution. In this system we use the different classifiers on this real time packets and KDD dataset for the comparison of obtained results. In this system we use the Data Pre-processing algorithm, Hybrid Feature Selection Algorithm and Modified Naïve Bayes Algorithm. Using these algorithms we improve the system accuracy and execution performance. After that we compare the SVM, CNN, KNN, ANN, and Proposed Naïve Bayes Classifiers to obtain the better results. We use LOIC DDoS assault generator instrument to make the assault at bundle getting time what's more use KDD compartment dataset for various assault pursue, for instance, DoS, R2L, L2R, Test and whatnot of correspondence orchestrate. The test results that proposed MNBIDS Strategy is progressively precise and valuable.