Research On Intrusion Detection Algorithm Based On Industrial Internet Traffic Analysis

At present,the industrial control system in our country's steel and petrochemical,highend equipment,rail transportation,medical facilities,power systems,nuclear facilities,and other fields are widely used,with the continuous development of informatization and industrialization in our country related technologies,and constantly fusion,frequent international industrial control system security attacks,industrial the safe operation of the Internet and national security,social stability and people's life and property safety are closely related.In recent years,China has focused on the development of industrial Internet platform,the integration of industry and network applications,and the development of industrial market in the international development potential.In the construction of industrial Internet,China's attention is not inferior to the developed countries,but in the aspect of information security,China is particularly inadequate.Therefore,it is urgent to study and strengthen the analysis of industrial Internet traffic,improve the security audit technology for industrial Internet security enterprises,so as to improve the safety monitoring and protection ability of the industrial Internet.Intrusion detection system can take the initiative to protect information security,to make up for the shortage of firewall,improve the industrial control system to identify abnormal behavior or virus attack early warning ability.With the continuous integration of industry and network,the real-time requirements of industrial Internet have been improved compared with the traditional Internet.The existing intrusion detection technology for industrial Internet traffic analysis has the problem of low detection efficiency,so it is necessary to design a reasonable intrusion detection method to improve the accuracy of intrusion detection.In order to solve the problem of low accuracy of intrusion detection technology of industrial Internet traffic detection,a CNN-SVM intrusion detection model was proposed by combining Support Vector Machine(SVM)and Convolutional Neural Networks(CNN)into the intrusion detection technology of industrial Internet.After network data is preprocessed,one-dimensional data is transformed into matrix form as input.Feature extraction and dimension reduction are carried out through the convolutional layer and the pooling layer.After dimensionalization reduction,the features are mapped to one dimension and handed to the improved SVM classifier for classification.The simulation experiment is carried out by using open industrial Internet data set.Experimental verification shows that CNN-SVM model can be applied in abnormal traffic detection of industrial Internet,and the accuracy rate reaches 92.47%.Compared with CNN and SVM intrusion detection models alone,the intrusion detection accuracy of industrial Internet traffic analysis is improved.