An Efficient Affinity Propagation Clustering Technique

Abstract

Clustering Problem is the important technique that has been addressed several times. It has importance for
data analysis in enormous application domains. There are numerous clustering methods available in the
literature, this paper addresses a Feature based Affinity Propagation (FBAP) clustering method that improves
the clustering solution. It increases intra cluster similarity by applying weighted feature to the data point before
decision making process in cluster assignment. Intrusion detection has been the point of attraction for many
researchers from last decades so the effectiveness of the algorithm is demonstrated on DARPA’s KDD Cup’99
training dataset.