Research on Network Intrusion Detection Method Based on Machine Learning and DBN Network
With the rapid development of computer networks,network intrusion has become more and more serious,and the traditional network intrusion detection methods have low detection efficiency and high false positive rate.In order to solve these problems,the research proposes an intrusion detection method based on support vector machine-deep belief network(SVM-DBN).By optimizing support vector machine(SVM),SVM-deep belief network(DBN)is fused.SVM,DBN and SVM-DBN are used to compare in network intrusion dataset.The results show that the SVM-DBN algorithm has the lowest error rate,which is 8.95％and 12.70％lower than the average of the error rates of DBN and SVM,respectively,and the SVM-DBN al-gorithm has a maximum absolute percentage error of 4.8％ when the number of training times is 140,which is better than the comparison methods.This indicates that SVM-DBN network can effectively improve the accuracy and efficiency of network in-trusion detection.
万方数据
维普
