Research On Network Intrusion Detection Methods Based On Feature Selection

With the rapid development of Internet technology,the Internet has penetrated into all areas of life.However,due to the existence of various vulnerabilities in the computer network system and the wide spread of a large number of hacker tools,network intrusions and attacks are emerging one after another,and traditional network intrusion detection methods have been unable to cope with the evolving new threats.In recent years,with the rapid development of machine learning technology,the use of machine learning methods for network intrusion detection has become a research hotspot.When conducting network intrusion detection,there are two main problems that need to be solved: The first is the problem of feature selection.Due to the high dimensionality and large quantity of network data,it consumes a lot of resources,and the existence of some redundant features will reduce the accuracy of classification.Therefore,it is necessary to perform feature selection on network data,delete redundant features,and reduce the number of features.The second is the problem of network intrusion detection.Network intrusion detection can be regarded as a classification problem,and a classification model of network data needs to be established to identify intrusion behaviors and normal behaviors.Therefore,this paper studies the feature selection problem and network intrusion detection problem based on the gradient boosting tree.The main work is as follows:Firstly,aiming at the problem of feature selection,this paper proposes and implements a feature selection method based on IBCO-GBDT.This method uses the Improved Border Collie Optimization algorithm as the search strategy,uses the Gradient Boosting Decision Tree as the classifier,and uses the number of features and classification accuracy as indicators of the evaluation function.The improved border collie optimization algorithm is used to package the gradient boosting tree for feature selection.Secondly,aiming at the problem of network intrusion detection,this paper proposes and implements a network intrusion detection method based on IWMA-XGBoost.This method uses e Xtreme Gradient Boosting tree to multi-classify network data.Since the parameters of XGBoost have a greater impact on the performance of the algorithm,the improved Woodpecker Mating Algorithm is used to optimize the hyperparameters of XGBoost.The organic integration of the above two methods provides a new solution for network intrusion detection.Finally,two network intrusion detection data sets KDD Cup 99 and UNSW-NB15 are used for experiments,and the feature selection method and network intrusion detection method proposed in this paper are compared with other feature selection methods and network intrusion detection methods.The methods are compared to comprehensively verify the effectiveness of the method proposed in this paper in the early network environment and the current network environment.The experimental results show that the feature selection method proposed in this paper can effectively reduce the number of features and improve the classification accuracy.The network intrusion detection method proposed in this paper can effectively improve the accuracy and recall rate of network intrusion detection classification,and reduce the false alarm rate.