A knowledge fusion based intrusion detection model

This paper presents a compound intrusion detection model based on a knowledge fusion algorithm. The proposed method suggests a set of related pattern knowledge objects with a well-defined structure, which instructs that pattern knowledge can be used for misuse detection or anomaly detection or both of them. Fusion algorithm is used to fuse original or real time knowledge objects to generate new useful ones and destroy those who may lead to miscar-riage of justice. In contrast with the conventional intrusion detection models, process in the paper combines anomaly and misuse detection so that it eliminates the flaws of a narrow definition for intrusion patterns and extends the known intrusions patterns to novel intrusions patterns. The experimental results show the feasibility of design rationale.