Research on Network Intrusion Detection Method Based on Improved NSA and CNN Algorithms
To understand conventional network intrusion problems,a network intrusion detection model based on Negative Selection Algorithm(NSA)is proposed.Considering the accuracy issue of the NSA model,a genetic model is introduced to optimize the detector,while reducing the dimensionality of the data to improve the detection effect.At the same time,a large-scale network intrusion model is constructed based on Convolutional Network Le-Net Lene(LeNet-5)for digital recognition,and syn-thetic data algorithm optimization is introduced.In conventional network intrusion detection,the classi-fication detection accuracy of the improved NSA model is 98.6％when the dataset is 20000,which is better than the other two models.In large-scale network intrusion detection,the proposed large-scale intrusion detection model has an accuracy of 0.912 in detecting U2R type datasets.It can be seen that the proposed network intrusion detection technology has good application effects and provides important technical references for network risk prevention.
万方数据
维普
