An application of the generalised K-means algorithm in decision-making processes

Hsin-Hung Wu, Jiunn I. Shieh, Anthony Y H Liao, Shih Yen Lin

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

A case study of applying the generalised K-means algorithm with different p values is provided to discuss the applicants' selection under a variety of criteria in an admission process. The properties of the generalised K-means algorithm are exploited in a decision-making process. When p is smaller and closer to zero, the results show the priorities are identical, which is to look for the applicants with even performance. In contrast, the most commonly used p values in K-means algorithm do not generate a systematic pattern. When p becomes larger and approaches ∞, the results show the priorities are difficult to tell, but the intention is to separate alternatives with a number of clusters, which is to look for the applicants with the greatest potential. Finally, in this case study, using smaller p values might provide stable priorities to select 21 applicants out of 36 participants.

Original languageEnglish
Pages (from-to)19-35
Number of pages17
JournalInternational Journal of Operational Research
Volume3
Issue number1-2
DOIs
Publication statusPublished - 2008 Apr 18

Fingerprint

K-means
Decision-making process
P value
Admission

All Science Journal Classification (ASJC) codes

  • Management Science and Operations Research

Cite this

Wu, Hsin-Hung ; Shieh, Jiunn I. ; Liao, Anthony Y H ; Lin, Shih Yen. / An application of the generalised K-means algorithm in decision-making processes. In: International Journal of Operational Research. 2008 ; Vol. 3, No. 1-2. pp. 19-35.
@article{ac67ae4797c54620ac88dbc59f336242,
title = "An application of the generalised K-means algorithm in decision-making processes",
abstract = "A case study of applying the generalised K-means algorithm with different p values is provided to discuss the applicants' selection under a variety of criteria in an admission process. The properties of the generalised K-means algorithm are exploited in a decision-making process. When p is smaller and closer to zero, the results show the priorities are identical, which is to look for the applicants with even performance. In contrast, the most commonly used p values in K-means algorithm do not generate a systematic pattern. When p becomes larger and approaches ∞, the results show the priorities are difficult to tell, but the intention is to separate alternatives with a number of clusters, which is to look for the applicants with the greatest potential. Finally, in this case study, using smaller p values might provide stable priorities to select 21 applicants out of 36 participants.",
author = "Hsin-Hung Wu and Shieh, {Jiunn I.} and Liao, {Anthony Y H} and Lin, {Shih Yen}",
year = "2008",
month = "4",
day = "18",
doi = "10.1504/IJOR.2008.016152",
language = "English",
volume = "3",
pages = "19--35",
journal = "International Journal of Operational Research",
issn = "1745-7645",
publisher = "Inderscience Enterprises Ltd",
number = "1-2",

}

An application of the generalised K-means algorithm in decision-making processes. / Wu, Hsin-Hung; Shieh, Jiunn I.; Liao, Anthony Y H; Lin, Shih Yen.

In: International Journal of Operational Research, Vol. 3, No. 1-2, 18.04.2008, p. 19-35.

Research output: Contribution to journalArticle

TY - JOUR

T1 - An application of the generalised K-means algorithm in decision-making processes

AU - Wu, Hsin-Hung

AU - Shieh, Jiunn I.

AU - Liao, Anthony Y H

AU - Lin, Shih Yen

PY - 2008/4/18

Y1 - 2008/4/18

N2 - A case study of applying the generalised K-means algorithm with different p values is provided to discuss the applicants' selection under a variety of criteria in an admission process. The properties of the generalised K-means algorithm are exploited in a decision-making process. When p is smaller and closer to zero, the results show the priorities are identical, which is to look for the applicants with even performance. In contrast, the most commonly used p values in K-means algorithm do not generate a systematic pattern. When p becomes larger and approaches ∞, the results show the priorities are difficult to tell, but the intention is to separate alternatives with a number of clusters, which is to look for the applicants with the greatest potential. Finally, in this case study, using smaller p values might provide stable priorities to select 21 applicants out of 36 participants.

AB - A case study of applying the generalised K-means algorithm with different p values is provided to discuss the applicants' selection under a variety of criteria in an admission process. The properties of the generalised K-means algorithm are exploited in a decision-making process. When p is smaller and closer to zero, the results show the priorities are identical, which is to look for the applicants with even performance. In contrast, the most commonly used p values in K-means algorithm do not generate a systematic pattern. When p becomes larger and approaches ∞, the results show the priorities are difficult to tell, but the intention is to separate alternatives with a number of clusters, which is to look for the applicants with the greatest potential. Finally, in this case study, using smaller p values might provide stable priorities to select 21 applicants out of 36 participants.

UR - http://www.scopus.com/inward/record.url?scp=42049100965&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=42049100965&partnerID=8YFLogxK

U2 - 10.1504/IJOR.2008.016152

DO - 10.1504/IJOR.2008.016152

M3 - Article

AN - SCOPUS:42049100965

VL - 3

SP - 19

EP - 35

JO - International Journal of Operational Research

JF - International Journal of Operational Research

SN - 1745-7645

IS - 1-2

ER -