The Research Of Intrusion Detection System Based On BP Neural Network

Because the network itself open and free features, resulting in some illegal molecular attacks, malicious destruction or infringement of network, security issues have become increasingly prominent. Constantly updated network attack methods and technology, make the security the traditional firewall, digital authentication and other protective measures have been unable to meet the demand of network security, intrusion detection technology emerge as the times require. However, due to limitations of intrusion detection algorithm, the current intrusion detection systems are still poor real-time performance, the false alarm rate is higher.Network intrusion detection model in the convergence rate and false negative rate of construction of traditional BP neural network is analyzed in this paper, the particle swarm algorithm applied to intrusion detection system; by attacking the principles of Probing and Dos, analysis of the characteristics of the Dos attack method, extracting characteristic data, to establish the feature set, a design the improvement of PSO intrusion detection model based on BP neural network and network intrusion based design, and on the model of the detection system, the simulation test proves that the improved effect of the system in the false alarm rate, convergence rate and false negative rate of. The main research work of this thesis are as the following:(1)The paper analyzing disadvantages of intrusion detection model is to construct the standard particle swarm optimization algorithm with BP neural network, through the introduction of the inertia weight factor, the shrinkage factor and multi object optimization strategy improved particle swarm optimization algorithm, and the improved particle swarm optimization and BP neural network.(2) It is used to design BP neural network using MATLAB tools, from training data and test data from the KDDCUP data sets, the neural network is trained.(3)The paper the trained BP neural network for intrusion detection, construction optimization in Intrusion Detection System Based on BP neural network, in order to improve the defense capability, through the linkage system and firewall, antivirus, anti spyware, establish the system protection system in all aspects, which have the ability of active defense system.In the end.The paper design the platform finally design experimental environment for performance analysis, improvement intrusion detection system based on PSO-BP neural network, proficiency testing system for detection of Probing attack and Dos attack, and compares it with traditional BP neural network. The experimental results show that the improved PSO-BP neural network, intrusion detection system can effectively to prevent from malicious attacks based on network, improve the efficiency and performance of detection, reduce the false negative rate and the false positive rate; it also proved that the improved PSO-BP neural network is feasible for intrusion detection.