An intrusion detection approach based on data mining

An intrusion is defined as any set of actions that compromise the integrity, confidentiality or availability of a resource. Data mining is to identify valid, novel, potentially useful, and ultimately understandable patterns in massive data. It is demanding to apply data mining techniques to detect various intrusions. This paper presents an approach to detect intrusion based on data mining frame work. In the framework, intrusion detection is thought of as clustering. The reduction algorithm is presented to cancel the redundant attribute set and obtain the optimal attribute set to form the input of the FCM. To find the reasonable initial centers easily, the advanced FCM is established, which improves the performance of intrusion detection since the traffic is large and the types of attack are various. In the illustrative example, the number of attributes is reduced greatly and the detection is in a high precision for the attacks of DoS and Probe, a low false positive rate in all types of attacks.