Network Intrusion Detection Based On IDBN-ELM

With the rapid development of Internet technology,the network environment is becoming increasingly complex and changeable,and the network security problem is becoming more and more serious.As one of the effective measures of network security defense,how to improve the detection efficiency of intrusion detection has become the research hotspot of the majority of researchers.In recent years,deep learning has made great achievements in various fields.Combining deep learning with intrusion detection technology to improve the efficiency of intrusion detection provides a new idea for traditional intrusion detection.Based on in-depth understanding of intrusion detection and deep learning theories and combining with the characteristics of KDD CUP99 data set,this paper chooses deep belief network and extreme learning machine as the research object of this paper.Aiming at the problem of low detection efficiency of current intrusion detection system,the following work is done:(1)Belief network data processing ability,in order to improve the depth in the depth of the belief network training phase,by putting on a layer of hidden features combined with the original data as the input of the next layer of hidden layer a layer of hidden layers on the training to improve the way is applied to intrusion detection system,strengthen the original data extraction ability,improve the network intrusion detection model of DBN.(2)In order to further improve the efficiency of intrusion detection,based on the first work and the good classification ability of the extreme learning machine,the paper proposes to replace the BP classifier in the original deep belief network with the extreme learning machine to construct the IDBN-ELM network intrusion detection model.(3)In order to verify the validity of the model proposed in this paper,the KDD CUP99 data set is used for simulation experiments.The experimental results show that the data extraction ability of the improved deep belief network can be effectively improved,and the combination of deep belief network and extreme learning machine can also improve the detection efficiency of intrusion detection.