A Approach Of Intrusion Detection Using Pattern Recognition

With the development of Internet, network security becomes more important than it has ever been. Intrusion detection technique, one of the key techniques used in network security, has also become the focus in the area of network security research. At present, intrusion detection technique is divided into anomaly detection and misuse detection. The measure of misuse detection is if the information to be dealt with matches the rule in the rule library based on known intrusive methods. Its advantage is the detective result is accurate, but it can only identify the known intrusive methods. Anomaly detection is an effort to recognize unusual behavior through a profile describing normal behavioral patterns. This measure can discriminate the unknown methods of intrusion but the technique is not mature and false alarm ratio is high.This paper creates two vector-sets through combining the advantages of the two measures. One describes the behavior of normal users and the other describes that of unusual users, and then introduces a pattern recognition algorithm which is used in image processing technique to judge the behavioral attribute of users to be detected.This paper primarily solves three questions as follow . the first is data source that the detection needed and preprocessing of these data; the second is introduction and analysis of the neighbor algorithm; the last is the realization of program.