Application Research Of Improved Whale Optimization Algorithm In Network Intrusion Detection

With the rapid development of the information age,not only the Internet's network structure is becoming more and more complex,but also the amount of information transmitted by the network is increasing.The resulting security threats have brought varying degrees of damage to social stability and economic development.Intrusion detection technology can detect attacks before cyber attacks cause extensive damage,which provides an important basis for the development of defense strategies and reduces the damage of network intrusions.In order to improve the effectiveness and classification performance of network intrusion detection,this thesis takes the research status of intrusion detection as the starting point,The problem of data redundancy caused by the huge number of data features in intrusion detection and the problem of low classification performance caused by improper selection of parameters in SVM classification model are studied.The main research work is as follows.(1)An Improved Binary Whale Optimization Algorithm(IBWOA)is proposed.When solving the optimization problem with high data feature dimension,the whale optimization algorithm has the disadvantages of slow convergence and easy to fall into local optimum.This thesis proposes an IBWOA algorithm to improve its convergence speed and the ability to jump out of local optimum by improving the convergence factor in the whale optimization algorithm and incorporating the particle swarm strategy into the update mechanism.(2)Intrusion detection feature selection method based on IBWOA algorithm.In order to reduce the feature redundancy caused by high data set dimension and large data volume during network intrusion detection,feature selection of data set is needed to reduce redundant features.Therefore,this thesis adopts the improved intrusion detection feature selection method of the IBWOA algorithm to extract features with large contribution in data set and delete redundant features to improve the effectiveness of network intrusion detection.(3)Intrusion detection method based on IBWOA algorithm for optimizing SVM.In order to improve the classification accuracy of network intrusion detection,it is necessary to optimize the parameters of SVM to find the optimal penalty factor and kernel function,and establish an optimal classification model.Therefore,this thesis uses the IBWOA algorithm to test the SVM parameter optimization problem to improve the classification accuracy of intrusion detection.Based on the feature selection method and SVM parameter optimization method of IBWOA algorithm in the thesis,it will be tested on multiple UCI and KDD CUP 99 data sets,and compared with genetic algorithm,particle swarm optimization algorithm and whale optimization algorithm.The experimental results show that the intrusion detection feature selection method based on IBWOA algorithm can select fewer feature subsets and the higher accuracy.The intrusion detection method based on IBWOA algorithm to optimize SVM can also achieve better classification performance.