Research And Implementation Of Network Intrusion Detection System Based On QBSO-ELM

In recent years,with the continuous development of global changes,the complex mix of the century’s epidemics,the international environment becoming increasingly complex and the global supply chain suffering shocks,the trend of confrontation in cyberspace has become increasingly obvious.Large-scale targeted cyber-attack activities are on the rise,and risks such as security breaches,data leaks and cyber-fraud are also appearing frequently,making the cyber-security situation increasingly critical.Intrusion detection plays a crucial role in cybersecurity,but traditional techniques do not perform well when dealing with high-dimensional and massive data from comprehensive information technology.Therefore,it has triggered researchers to apply deep learning to intrusion detection,and certain results have been achieved,but the models still suffer from problems such as overfitting,low detection rates,high false alarm rates and poor results in detecting a few classes of samples.The main research work in this paper is as follows.In this paper,an extreme learning machine model based on quantum swarm intelligence optimization algorithm is proposed,aiming to improve the accuracy of intrusion detection and reduce the false alarm rate.Firstly,the LSQR algorithm is applied to optimize the extreme learning machine model,which effectively reduces the computational effort in solving.Secondly,the quantum beetle swarm optimization(QBSO)algorithm is proposed to enable individual beetle to learn group experience and their own experience,which improves the convergence speed while moving purposefully.Finally,the experimental results show that the method effectively improves the accuracy of intrusion detection.To solve the problem of imbalance in intrusion detection dataset,a network intrusion detection model based on Focal loss function and Convolutional Neural Networks(CNN)is proposed.In the data processing stage,CNN is used to extract features,which can automatically extract advanced features from the input data that are more effective for the intrusion detection and thus improve the classification accuracy.At the same time,in order to enhance the effectiveness of the model for multiple classification,Adaptive Synthetic Sampling(ADASYN)is used to make the sample data distribution basically balanced,and the Focal loss loss function is used to calculate the loss with different weights,so that the network can better learn the features of different classes of samples,thus effectively reducing the negative impact of the unbalanced data set on the classifier and improving the effect of minority class sample detection.The experimental results show that the method can effectively improve the performance of the classification model.Finally,based on Flask and React and combined with the algorithm model proposed in this paper,an intrusion detection system is developed,which implements the basic management functions and intrusion detection functions,providing a new solution for securing the network.