The Research Of Intrusion Detection Based On Deep Learning

With the continuous development of Internet technology and the rapid spread of network applications,people's daily life and work are more and more dependent on the Internet.However,while the Internet brings convenience to people,it also creates some security problems,such as various viruses?vulnerabilities and attacks,which can cause economic losses and have a bad influence on the sustainable and steady development of society.Intrusion detection is an effective network security defense technology by collecting and analyzing network information data to identify intrusions in computer network systems.The use of traditional machine learning methods to deal with large-scale intrusion data can easily lead to problems,such as low detection accuracy and slow detection speed.In order to solve these problems,this paper proposes a deep learning based hybrid intrusion detection model based on the deep analysis of deep learning model and intrusion detection method.The main work and innovation of this article is as follows:(1)Faced with large-scale network data streams,how to quickly select effective features from them as the basis for intrusion recognition is a key research challenge in the field of intrusion detection.In this article,we will apply deep learning algorithm to feature learning of intrusion detection,and propose a feature learning algorithm based on DBN.The algorithm can map non-linear,high-dimensional raw data to low-dimensional space to reduce the feature dimension of the data set and obtain the optimal low-dimensional representation of the original data set.Compared with existing algorithms,this algorithm has better feature representation capability and improves the accuracy of intrusion detection.(2)Intrusion detection classification algorithms based on traditional support vector machines have low training speed when faced with large-scale data and cannot identify specific attack types.This paper synthesizes the advantages of and Partial Binary Tree(PBT)and Twin Support Vector Machines(TSVM),and constructs a PBT-TSVM algorithm based on partial binary tree for twin support vector machines,which can identify network intrusion data.Finally,experiments show that the PBT-TSVM algorithm proposed in this paper not only improves the detection accuracy,but also reduces the false alarm rate to a certain extent.It also has good performance in training time.(3)By studying the basic intrusion detection model,we propose a hybrid intrusion detection model DBN-PBT-TSVM based on DBN.Firstly,the network data is preprocessed to obtain the standard data set.Then,it is trained and reduced dimension by DBN-based feature learning algorithm.Finally,the classification and identification of intrusion data is implemented by the algorithm PBT-TSVM.Compared with traditional intrusion detection algorithms,DBN-PBT-TSVM model not only improves the accuracy of classification,but also significantly improves the detection speed,especially when dealing with large-scale data.