Detection of intrusion using evolutionary soft computing techniques

An intrusion detection system has the ability to detect known as well as unknown attacks. Conventionally, IDS uses association rules and simple partitioning scheme to handle quantitative data. Fuzzy partitioning scheme found to overcome the problem of vagueness of boundary and association rules are also replaced with fuzzy class association rules. Evolutionary algorithms do global search in order to discover new interesting classification rules. GNP is found better than other evolutionary algorithms due to its structure. This paper describes how evolutionary computing technologies are better to build a classifier with minimum number of class association rules and maximizing accuracy of classified patterns in intrusion detection problem.