Research On The Application Of Data Mining In The Intrusion Detection

Since the conception of Intrusion Detection was brought forward in 1980s, it has become to an essential technology to constitute the integrated modern network security during 20 years' development. With the improvement of the speed of the network, more and more network data can be produced, so it is unrealistic to find the intrusion patterns by security experts' eyes and intuition. Therefore, Wenkee lee introduced technology of data mining into intrusion detection in order to build intrusion detection system automatically.Data mining-based IDSs have lower detection performance than traditional handcrafted signature based methods. In order to improve the detection performance of data mining-based intrusion detection system, we apply combination multiple classifications by fuzzy integral into intrusion detection and bring forward the method of Multiple Decision Tree Fusion using Fuzzy Integral (FIFDT) and the method of Multiple Neural Network Fusion using Fuzzy Integral (MNNF) in this paper. Using KDD99 as experiment data, the experiment results show that these methods can improve intrusion detection performance further. In this paper, firstly, the knowledge of intrusion detection and the knowledge of data mining are introduced; secondly, fusion methods referred to in this paper are introduced, and finally, the basic ideas, experiment methods and experiment results of the model of FIFDT and the model of MNNF are introduced in stress.