Network Intrusion Detection Based On Fuzzy-neural Hybrid System

With the rapid development of the network technology, the Internet is widely usedin People’s life. Meanwhile, the network has been attackers’ main goals. There are alarge number of attacks on the network and the direct loss is being more and moreserious.Presently, the internet security use the are static security technology mainly, itplays a certain role to preventing the illegal intrusion, but when the attacks happens,especially when the new attack methods are used, the system would suffer anunexpected loss. Therefore, we need active and dynamic security technology. TheIntrusion Detection System (IDS) is a new dynamic security technology whichappeared20years ago. The IDS can detect many kinds of intrusion actions, it’s animportant part of current computer network security system.The Artificial Neural Networks intrusion detection has a greater advantage thantraditional intrusion detection. The Artificial Neural Networks intrusion detectionsystem can find the rules from a large number of the unknown modal complex data,it can construct the relational model which has a good effect on categories andforecast.All of these things are done through continuous learning and training. In thispaper, Fuzzy-Neural Hybrid System (FNHS) is used to be a intrusion detecting method,it can improve the accuracy and warning rate effectively, in the Intrusion DetectionSystem, we propose an effective method to reduce the input dimension, then thenetwork’s size will be reduced and the training speed will be improved, TheKDDCup99dataset is used to be the training and the testing date in our experiment,and the good effect is shown, so it’s Suitable for The Intrusion Detection System...