Comprehensive analysis of data clustering algorithms

Abstract

We have given a comprehensive comparative analysis of various clustering algorithms. Clustering algorithms usually employ distance metric or similarity matrix to cluster the data set into different partitions. Well known clustering algorithms have been widely used in various disciplines. Type of clustering algorithm used depends upon the application and data set used in that field. Numerical data set is comparatively easy to implement as data are invariably real numbers. Others type of data set such as categorical, time series, boolean, and spatial, temporal have limited applications. It is observed that there is no optimal solution for handling problems with large data sets of mixed and categorical attributes. Some of the algorithms can be applied but their performance degrades as the size of data keeps on increasing. © 2013 Springer Science+Business Media Dordrecht.