Intrusion Detection Method Based On JRNB Algorithm

With the rapid development of the mobile Internet,the popularity of the Internet is getting higher and higher,which has brought a huge impact on people’s lives,ranging from the global political and economic environment to the personal work life.While people enjoy the convenience brought by the network,the accompanying network security incidents are becoming more frequent,such as malicious attacks by hackers and Trojan horses.These hidden network security risks may lead to information leakage,property loss,and serious It will endanger national security.Therefore,people pay more and more attention to network and information security.Intrusion detection is a technology that proactively defends against harmful behaviors that may occur in the network.As an important guarantee of information security,it can effectively detect possible attack behaviors,so that corresponding safeguard measures can be taken to avoid causing damage to the network.Based on the current research status of intrusion detection and the results of previous studies,this paper proposes an intrusion detection method based on feature weighted improved Naive Bayes algorithm for existing problems.This method is an extension of intrusion detection technology.The main work and innovation of this article are as follows:(1)By analyzing the problems existing in the traditional NB algorithm,the JS divergence and inverse category frequency(RCF)are introduced to measure the weight of each feature item,so as to compensate for the lack of equal analysis of all feature items in the NB algorithm,thereby reducing impact of conditional independence assumptions.(2)In order to reduce redundant attributes and improve the independence between attributes,singular value decomposition(SVD)and principal component analysis(PCA)are used to reduce the dimensionality of the training sample data and construct a new attribute set,and then on the optimized data set,building an improved naive Bayes classifier to achieve the purpose of improving classification accuracy.(3)In intrusion detection,the above-mentioned related data processing methods and weighted Naive Bayes(JRNB)algorithm based on JS divergence and anti-class frequency are used to improve the detection effect.Finally,the NSL-KDD data set is used for simulation experiments.This data set is an optimized version of the KDDCUP′99 data set,which makes the experimental evaluation more effective.The analysis of the experimental results verifies the effectiveness of the JRNB algorithm proposed in this paper in intrusion detection.Compared with other similar algorithms,it also shows certain advantages.