Data Mining In Intrusion Detection Applications

With the rapid development of computer networks, people's life, work, learning has become increasingly inseparable from the computer network. People come to realize that while the development of computer networks, security problems caused by more serious harm. Computer network security is one of an attacker using the victim's computer, computer network intrusion, unauthorized access to system resources, modify data on victims, victims and computer operations, to steal, destroy, control and other related illegal activities. Growing network intrusions, making the social and academic network intrusion detection for more and more attention, how quickly and accurately detect network intrusion network security has become a major research topic.Intrusion Detection System (IDS) can be used to identify the computer network intrusion. This paper studies the characteristics of intrusion detection data, designed based on fuzzy data mining intrusion detection system.One of the intrusion detection system is the use of fuzzy support vector machine active learning, access classifier for intrusion detection. Rough set support vector machine classifier construction method, by rough set to reduce redundant data is about to accelerate the learning process. The method of false positives is low, but the omission rate.Another intrusion detection system is based on the fuzzy frequent pattern growth (Frequent Pattern growth, FP-growth) method, the core of association rules engine, with a set of fuzzy association rules to describe the different categories, estimated to be detected by various types of samples and set of matching rules in order to best match the type of a sample of the test results. As Fuzzy Apriori algorithm efficiency is not high enough, this paper fuzzy FP-growth algorithm, the FP-tree fuzzy construction and mining process, and propose a new method for fuzzy FP-tree pruning, removing not included in the rules of entry, speed up the mining process; also directly derived from the FP-tree fuzzy association rules, replacing the final scan the database for association rules derived steps to accelerate the entire training process. The technology theory and the fuzzy association rule mining frequent patterns combined results show that the technology to effectively improve the learning efficiency and reduce the false negative rate.This article mainly focused on data mining based on fuzzy intrusion detection application. In response to these two kinds of data mining algorithms to explore the fuzzy data mining method applied to intrusion detection.