Mining of Network Data for Intrusion Detection using Multi-Dimensional Hierarchical K Means Clustering employed with Hybrid ABC-DT

摘要

The interest of Content centric network (CCN) increase tremendously because of its application as a future internet. Since the challenges in CCN also increases with the security and privacy attacks present in the network. An efficient and viable security instrument is required to verify substance and guard against obscure and new types of assaults and inconsistencies. As a rule, clustering calculations would fit the necessities for building a decent Intrusion recognition framework. The customary calculations experience the nearby combination and affectability to determination of the cluster centroids. In this article, we present a narrative multi-dimensional Hierarchical k means-clustering algorithm for Intrusion recognition system. Initially the clustering algorithm is projected to form a numeral of clusters in the CCN and then the optimal clusters are selected by the utilization of Cuckoo search optimization algorithm (CSOA). Finally, we employ an Artificial Bee colony based Decision Tree classifier in order to categorize the customary and anomalous cases present in the network by means of the extract features. The anticipated technique will be implemented on MATLAB working platform and tested on widely used KDD CUP 99 dataset. The consequences of the above implementation will be compared through existing methods. Trial results exhibit that the proposed calculation can accomplish to the ideal number of clusters, well-isolated groups, just as increment the high discovery rate and abatement the bogus positive rate simultaneously when contrasted with some other surely understood clustering calculations.