Research On Network Intrusion Detection Based On Pattern Recognition Algorithm

With the popularity of internet, internet affects the polity, economy, culture, military and life. In order to share and use of the network information and resources, more and more people, corporations and government departments connect computers with the internet. Therefore, people rely on the network increasingly. With increasing connectivity between computers, the network also produces a wide range of issues. The internet security becomes one of the key problems in the world. According to the statistics, the global number of virus has been more than 40,000. The average network intrusion incident occurs every twenty seconds. Each year the global loss caused by the issue of security can be calculated in the magnitude of one trillion U.S. dollars. Information security directly influences the interests of nation, corporation and individual. Traditionally, firewall is widely used security measures. Nowadays, the tools and tactics turn more complex and diverse. A simple firewall policy has been unable to meet people's needs. Therefore, the protection of computer systems, network systems and the security of information infrastructure have to be addressed immediately. Intrusion detection technology is an important part of internet security.This thesis proposes a novel intrusion detection system, which combines the supervised classifiers and unsupervised clustering to detect intrusions. Decision Tree, Na(?)ve Bayes and Bayesian clustering are used at different levels. We also have made improvements to the Naive Bayes algorithm by choosing different attributes for different classes. The experiments demonstrate the effectiveness of the proposed approach, especially for U2R and R2L type attacks. The detection rate is significantly improved.