Pattern Recognition using the Fuzzy c-means Technique

In the field of pattern recognition due to the fundamental involvement of human perception and inadequacy of standard Mathematics to deal with its complex and ambiguously defined system, different fuzzy techniques have been applied as an appropriate alternative. A pattern recognition system has to undergo basically the steps of preprocessing, feature extraction and selection, classifier design and optimization. In our work the data we have analyzed is in the form of numerical vectors, with a number of clusters predefined. Therefore the fuzzy c-means technique of Bezdek has been considered for our work. Although in the fuzzy c-means technique Euclidean distance has been used to obtain the membership values of the objects in different clusters, in our present work along with Euclidean distance we have used other distances like Canberra distance, Hamming distance to see the differences in outputs.