Cluster Analysis with Various Algorithms for Mixed Data

Cluster Analysis with Various Algorithms for Mixed Data

Abha Sharma (Maulana Azad National Institute of Technology, India) and R. S. Thakur (Maulana Azad National Institute of Technology, India)
Copyright: © 2017 |Pages: 36
DOI: 10.4018/978-1-5225-0536-5.ch014
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Analyzing clustering of mixed data set is a complex problem. Very useful clustering algorithms like k-means, fuzzy c-means, hierarchical methods etc. developed to extract hidden groups from numeric data. In this paper, the mixed data is converted into pure numeric with a conversion method, the various algorithm of numeric data has been applied on various well known mixed datasets, to exploit the inherent structure of the mixed data. Experimental results shows how smoothly the mixed data is giving better results on universally applicable clustering algorithms for numeric data.
Chapter Preview
Top

Background

The k-means algorithm which is the base of all the clustering algorithms and implemented in various areas. But the algorithm has major drawback of not handling the type of data other than numeric. Huang (1997) proposed k-prototype algorithm which combines the k-modes and the k-means algorithm using two types of distance measures Euclidian distance and match mismatch measure respectively, therefore it can handle mixed numeric and categorical datasets.

Complete Chapter List

Search this Book:
Reset