A New Concept of Network Intrusion Detection using Fuzzy Clustering
Richa Sampat, Shilpa S. Sonawani · 2015
Internet has become a vital part of any organization. Sensitive and confidential information is being sent over the network. But with the growth of internet, intrusion and attacks have also increased. Thus, there arises a need of robust and powerful intrusion detection systems which can detect the attacks. Recently, many novel methods are experimented to build strong IDSs. The aim of this paper is to present a methodology that can recognise and detect attacks efficiently. In this paper, we implement the FCM algorithm and successfully integrate it in WEKA to expand the system functions of the open-source platform, so that users can directly call the FCM algorithm to do fuzzy clustering analysis. Besides, considering the shortcoming of the classical FCM algorithm in selecting the initial cluster centers, we represent an improved FCM algorithm which adopts a new strategy to optimize the selection of original cluster centers. A novel classification via dynamic fuzzy c means clustering algorithm has been proposed to build an efficient anomaly based network intrusion detection model. A subset of KDDCup 1999 intrusion detection benchmark dataset has been used for the experiment. The proposed novel concept will be efficient in terms of detection accuracy, low false positive rate in comparison to the other existing methods.