The purpose of Intrusion Detection System is to detect activities contrary to the security rules of system or threat to system security. Traditional Intrusion Detection Systems are largely based on the use of expert systems or statistical methods. These detection models depend on empirical data, lacking of accuracy and the capability to detect new offensives. In this paper, the improved k-means clustering algorithm is applied into Intrusion Detection System, and this system can detect the type of data automatically, without experience, without guidance. Experiments show the effectiveness of th...