Optimal variable weighting for hierarchical clustering: An alternating least-squares algorithm

Geert De Soete, Wayne Desarbo, J. Douglas Carroll

Research output: Contribution to journalArticle

30 Citations (Scopus)

Abstract

This paper presents the development of a new methodology which simultaneously estimates in a least-squares fashion both an ultrametric tree and respective variable weightings for profile data that have been converted into (weighted) Euclidean distances. We first review the relevant classification literature on this topic. The new methodology is presented including the alternating least-squares algorithm used to estimate the parameters. The method is applied to a synthetic data set with known structure as a test of its operation. An application of this new methodology to ethnic group rating data is also discussed. Finally, extensions of the procedure to model additive, multiple, and three-way trees are mentioned.

Original languageEnglish (US)
Pages (from-to)173-192
Number of pages20
JournalJournal of Classification
Volume2
Issue number1
DOIs
StatePublished - Dec 1 1985

Fingerprint

Alternating Least Squares
Least Square Algorithm
Hierarchical Clustering
weighting
Least-Squares Analysis
Weighting
Cluster Analysis
Methodology
methodology
Ethnic Groups
Additive Models
Synthetic Data
Euclidean Distance
Estimate
Least Squares
ethnic group
rating
Hierarchical clustering
Least squares
Datasets

All Science Journal Classification (ASJC) codes

  • Mathematics (miscellaneous)
  • Psychology (miscellaneous)
  • Statistics, Probability and Uncertainty
  • Library and Information Sciences

Cite this

Soete, Geert De ; Desarbo, Wayne ; Carroll, J. Douglas. / Optimal variable weighting for hierarchical clustering : An alternating least-squares algorithm. In: Journal of Classification. 1985 ; Vol. 2, No. 1. pp. 173-192.
@article{674b10a87ab340188544facb0805cbdd,
title = "Optimal variable weighting for hierarchical clustering: An alternating least-squares algorithm",
abstract = "This paper presents the development of a new methodology which simultaneously estimates in a least-squares fashion both an ultrametric tree and respective variable weightings for profile data that have been converted into (weighted) Euclidean distances. We first review the relevant classification literature on this topic. The new methodology is presented including the alternating least-squares algorithm used to estimate the parameters. The method is applied to a synthetic data set with known structure as a test of its operation. An application of this new methodology to ethnic group rating data is also discussed. Finally, extensions of the procedure to model additive, multiple, and three-way trees are mentioned.",
author = "Soete, {Geert De} and Wayne Desarbo and Carroll, {J. Douglas}",
year = "1985",
month = "12",
day = "1",
doi = "10.1007/BF01908074",
language = "English (US)",
volume = "2",
pages = "173--192",
journal = "Journal of Classification",
issn = "0176-4268",
publisher = "Springer New York",
number = "1",

}

Optimal variable weighting for hierarchical clustering : An alternating least-squares algorithm. / Soete, Geert De; Desarbo, Wayne; Carroll, J. Douglas.

In: Journal of Classification, Vol. 2, No. 1, 01.12.1985, p. 173-192.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Optimal variable weighting for hierarchical clustering

T2 - An alternating least-squares algorithm

AU - Soete, Geert De

AU - Desarbo, Wayne

AU - Carroll, J. Douglas

PY - 1985/12/1

Y1 - 1985/12/1

N2 - This paper presents the development of a new methodology which simultaneously estimates in a least-squares fashion both an ultrametric tree and respective variable weightings for profile data that have been converted into (weighted) Euclidean distances. We first review the relevant classification literature on this topic. The new methodology is presented including the alternating least-squares algorithm used to estimate the parameters. The method is applied to a synthetic data set with known structure as a test of its operation. An application of this new methodology to ethnic group rating data is also discussed. Finally, extensions of the procedure to model additive, multiple, and three-way trees are mentioned.

AB - This paper presents the development of a new methodology which simultaneously estimates in a least-squares fashion both an ultrametric tree and respective variable weightings for profile data that have been converted into (weighted) Euclidean distances. We first review the relevant classification literature on this topic. The new methodology is presented including the alternating least-squares algorithm used to estimate the parameters. The method is applied to a synthetic data set with known structure as a test of its operation. An application of this new methodology to ethnic group rating data is also discussed. Finally, extensions of the procedure to model additive, multiple, and three-way trees are mentioned.

UR - http://www.scopus.com/inward/record.url?scp=0001332591&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0001332591&partnerID=8YFLogxK

U2 - 10.1007/BF01908074

DO - 10.1007/BF01908074

M3 - Article

AN - SCOPUS:0001332591

VL - 2

SP - 173

EP - 192

JO - Journal of Classification

JF - Journal of Classification

SN - 0176-4268

IS - 1

ER -