Psychometrika

, Volume 66, Issue 2, pp 249–270

A variable-selection heuristic for K-means clustering

Authors

    • Marketing Department, College of BusinessFlorida State University
  • J. Dennis Cradit
    • Marketing Department, College of BusinessFlorida State University
Articles

DOI: 10.1007/BF02294838

Cite this article as:
Brusco, M.J. & Cradit, J.D. Psychometrika (2001) 66: 249. doi:10.1007/BF02294838

Abstract

One of the most vexing problems in cluster analysis is the selection and/or weighting of variables in order to include those that truly define cluster structure, while eliminating those that might mask such structure. This paper presents a variable-selection heuristic for nonhierarchical (K-means) cluster analysis based on the adjusted Rand index for measuring cluster recovery. The heuristic was subjected to Monte Carlo testing across more than 2200 datasets with known cluster structure. The results indicate the heuristic is extremely effective at eliminating masking variables. A cluster analysis of real-world financial services data revealed that using the variable-selection heuristic prior to the K-means algorithm resulted in greater cluster stability.

Key words

cluster analysisK-means partitioningvariable selectionheuristics
Download to read the full article text

Copyright information

© The Psychometric Society 2001