Research And Application Of Optimized Sparrow Search Algorithm In Intrusion Detection

With the development of 5G technology,the internet penetration rate in China is constantly increasing,and the number of internet users continues to increase.The internet is also promoting the development of other industries,such as express,takeout and e-commerce,providing people with more convenient and efficient services.However,it is followed by the continuous intensification of network security problems,a variety of unknown network threats continue to emerge,which makes the development of the internet facing a huge challenge.Intrusion detection is a technology widely used in network security.It can effectively identify and prevent network threats,so as to ensure network security and maintain the normal operation and stable environment of network.Therefore,this paper proposes a binary intrusion detection model and a multi-label classification intrusion detection model.In this paper,a binary intrusion detection model based on optimized sparrow search algorithm is proposed to solve the binary intrusion detection task.To alleviate the shortcomings of BP neural network,sparrow search algorithm is used to search for the optimal initial weights and biases.In this paper,the logistic-tent chaotic mapping is used to initialize sparrow population and generate spatially evenly distributed individuals,so as to improve the population diversity of sparrow search algorithm.Aiming at the problem of decreasing population diversity of sparrow search algorithm in the late iteration period,this paper uses genetic algorithm to improve sparrow search algorithm.By controlling the selection,crossover and mutation strategies of the genetic algorithm,sparrow search algorithm can increase population diversity in the late iteration period.In this paper,the proposed intrusion detection model is compared with support vector machines,logistic regression,decision trees,random forests and other algorithms proposed in paper.The experimental results show that the intrusion detection model proposed in this paper has a higher recall in the task of intrusion detection,and has better performance and application prospects.Aiming at the multi-label classification task of intrusion detection,this paper proposes a multi-label classification intrusion detection model based on classifier chains.In order to improve the testing ability of the model for minority categories,Borderline SMOTE algorithm is used in this paper to double oversampling for minority categories and increase the diversity of data set.Classifier chains model is an effective method to solve the problem of multi-label classification.It can identify the correlation between labels and deal with the dependency between different labels.In the multi-label classification intrusion detection model,sparrow search algorithm based on genetic algorithm optimization is used to search the structure of the classifier chains,so as to optimize the model performance.Experimental results show that the average detection rate of the multi-label classification intrusion detection model proposed in this paper is higher than that of other models,but there is still room for improvement.This model provides an effective way of thinking and implementation for multi-label classification intrusion detection.Finally,this paper designs a visual intrusion detection system,which aims to provide intuitive and comprehensive operation interface and functional support for managers.The front end of the system mainly uses Vue and Element technology,and the back end chooses Django framework,which makes the combination of the system and intrusion detection model more convenient.The system can show the training process and results,and present the experimental data in the way of visual charts,which can help administrators to better understand the training situation of the model.The intrusion detection system developed in this paper provides a new idea,which can effectively apply the intrusion detection model to the actual scene and has certain practical significance.