The Implement And Research Of The Intrusion Detection System Technology Based On Date Mining And Plan Recognition

As network-based computer systems play increasingly vital roles in modern society, to take the place of traditional security program, Intrusion Detection becomes a new generation to an organizational information security program. As a key factor of information security program, Intrusion Detection (ID) provides critical protections from potential attempts to exploit computer resource vulnerabilities which greatly fetched up the limitation of access controlling and identity authentication. So the application of intelligent technology in intrusion detection system rises recently.Based on the former research of ID, this paper designs and presents an intelligent network-based intrusion detection system model. The paper introduces an overview of network security and development trend and direction of Intrusion Detection System. Then the paper offers a total design method, that is, to combine the data mining based Intrusion Detection with plan recognition Intrusion Forecasting. IDS collects data on multiple points, exchanges information via a uniform network interface, detects uncertain behavior by data mining, and analyzes intrusion from outside and activities without authorization from inside. Meanwhile, the model combines plan recognition to identify the intention of intrusion and predicts the future actions of attackers and responds to their appropriate actions. This paper mainly focuses on the data accessing, data mining, repository matching and plan recognition. During the detection, the paper runs the whole process of data mining through the ID technology, and gathers every data mining algorithm in a uniform frame to construct a system that can be accepted by different kinds of user. In the end, the paper summarizes the research the author does, and points out some steps to do next.