The Intrusion Detection System Based On Back Propagation Neural Network

Along with the increasing of the network security risk coefficient,the firewall as the most important security measurecan not satisfy the people's need to network security,As a useful complement to the firewall,IDS can help the network system find the attack quickly. It enlarged the security management capabilities of system administrator,(include Security audit,watch,attack identify and response) It improved the integrality of in formation security cadre.With the development of the network security technology,the application of the IDS in the network environment is more and more common.For the work of the enterprises,governments and other organizations has been depended on the computer network system more and more,The urgent affairs is to protect the system from attacking. IDS has become an important and indispensable means.In order to overcome the default of the IDS models,we must find an appropriate way to solve the problem.This paper introduces the basic concepts of intrusion detection,analyzes intrusion detection technique commonly in existing IDS model.We discuss the strengths and weaknesses of these methods,we focus on advantages and practical significance of the neural network-based intrusion detection methods.This paper gives a detailed describing to the deducing of BP algorithm and its betterment arithmetic of Levenberg-Marquardt(LM) optimized algorithm.This paper introduces the neural network technology in IDS model, and puts forward a detailed design scheme of intrusion detection model based on neural network.Finally,according to the BP algorithm defects,using the MATLAB language to achieve BP algorithm's L-M optimization,and using MATLAB simulation to prove the effectiveness of the system.It reduces developers'difficulties in algorithm'realizing, there is a certain degree of value in practice.