Binary Information Press
The intrusion detection is one kind of technology, which can detect any intrusion behavior that attempts to destroy resources' integrity, security, usability and so on, and adopts the countermeasure. And k-means is commonly used method for intrusion detection clustering algorithm. Its clustering results depend on the distance measurement choice. The traditional k-means algorithm usually take the Euclidean distance reciprocal as a similarity measurement, while this distance usually depends on all attributes in the same extension. When the k-means algorithm is adopted in intrusion detection and the importance degree of each attribute is definitely different. But depending on the similarity measurement of all attributes in the same degree tend to arouse misleading.