Data clustering, also called data segmentation, aims to partition a collection of data into a predefined number of subsets (or clusters) that are optimal in terms of some predefined criterion function. Data clustering is a fundamental and enabling tool that has a broad range of applications in many areas. Because of this, research on data clustering techniques has been the focus of considerable attention from multidisciplinary research communities such as pattern recognition, machine learning, data mining, information retrieval, bio-informatics, etc.

Generally speaking, to develop a data clustering method one needs to address the following three basic problems:
  1. 1.

    What is the model for modeling (or what is the assumption for the distribution of ) the given data set?

  2. 2.

    What is the criterion function to be optimized by the clustering process?

  3. 3.

    What is the computation algorithm for carrying out the optimization?


The data model together with the criterion function determine the data clustering capability, while the computation algorithm determines how effectively the designated clustering result can be obtained. Good clustering results are those that correspond well to human perceptions, and such results should be obtained by using computationally effective algorithms. Therefore, a good data clustering method can be defined as the one that constantly produces clustering results that correspond well to human perceptions, using a computationally effective algorithm.


Cluster Technique Spectral Cluster Data Cluster Document Cluster Document Corpus 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer Science+Business Media, LLC 2007

Personalised recommendations