A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients

Md Atiqul Islam; Behrooz Z. Alizadeh; Edwin R. van den Heuvel; Richard Bruggeman; Wiepke Cahn; Lieuwe de Haan; René S. Kahn; Carin Meijer; Inez Myin-Germeys; Jim van Os; Durk Wiersma

doi:10.1080/23737484.2015.1103670

A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients

Md Atiqul Islam^*, Behrooz Z. Alizadeh, Edwin R. van den Heuvel, Richard Bruggeman, Wiepke Cahn, Lieuwe de Haan, René S. Kahn, Carin Meijer, Inez Myin-Germeys, Jim van Os, Durk Wiersma

^*Corresponding author for this work

Research output: Contribution to journal › Article › Academic › peer-review

7 Citations (Scopus)

Abstract

Finding clusters in a complex dataset is not straightforward. Different indices were developed to quantify the number of clusters. Their performances were studied using unrealistic simulations, since they were considered at low dimensions. We investigated 14 indices for eight-dimensional data using simulations based on cognition measures. We focused on hierarchical clustering with Ward’s agglomerative technique. Results indicated that Duda and Hart, Hartigan and Gap/pc were best performing. They estimated the number of clusters within ±1 with high probabilities. Duda and Hart index was most consistent, while Gap/pc and WGap/pc together made a good distinction between single and multiple clusters.

Original language	English
Pages (from-to)	98-113
Number of pages	16
Journal	Communications in Statistics Case Studies Data Analysis and Applications
Volume	1
Issue number	2
DOIs	https://doi.org/10.1080/23737484.2015.1103670
Publication status	Published - 3 Apr 2015

Keywords

Cluster analysis
cluster indices
hierarchical clustering
homogeneous subgroups
number of clusters

Access to Document

10.1080/23737484.2015.1103670

Cite this

Islam, M. A., Alizadeh, B. Z., van den Heuvel, E. R., Bruggeman, R., Cahn, W., de Haan, L., Kahn, R. S., Meijer, C., Myin-Germeys, I., van Os, J., & Wiersma, D. (2015). A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients. Communications in Statistics Case Studies Data Analysis and Applications, 1(2), 98-113. https://doi.org/10.1080/23737484.2015.1103670

@article{95c3a2f4d9804da483ee433fb0a570c1,

title = "A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients",

abstract = "Finding clusters in a complex dataset is not straightforward. Different indices were developed to quantify the number of clusters. Their performances were studied using unrealistic simulations, since they were considered at low dimensions. We investigated 14 indices for eight-dimensional data using simulations based on cognition measures. We focused on hierarchical clustering with Ward{\textquoteright}s agglomerative technique. Results indicated that Duda and Hart, Hartigan and Gap/pc were best performing. They estimated the number of clusters within ±1 with high probabilities. Duda and Hart index was most consistent, while Gap/pc and WGap/pc together made a good distinction between single and multiple clusters.",

keywords = "Cluster analysis, cluster indices, hierarchical clustering, homogeneous subgroups, number of clusters",

author = "Islam, {Md Atiqul} and Alizadeh, {Behrooz Z.} and {van den Heuvel}, {Edwin R.} and Richard Bruggeman and Wiepke Cahn and {de Haan}, Lieuwe and Kahn, {Ren{\'e} S.} and Carin Meijer and Inez Myin-Germeys and {van Os}, Jim and Durk Wiersma",

year = "2015",

month = apr,

day = "3",

doi = "10.1080/23737484.2015.1103670",

language = "English",

volume = "1",

pages = "98--113",

number = "2",

}

Islam, MA, Alizadeh, BZ, van den Heuvel, ER, Bruggeman, R, Cahn, W, de Haan, L, Kahn, RS, Meijer, C, Myin-Germeys, I, van Os, J & Wiersma, D 2015, 'A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients', Communications in Statistics Case Studies Data Analysis and Applications, vol. 1, no. 2, pp. 98-113. https://doi.org/10.1080/23737484.2015.1103670

A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients. / Islam, Md Atiqul; Alizadeh, Behrooz Z.; van den Heuvel, Edwin R. et al.
In: Communications in Statistics Case Studies Data Analysis and Applications, Vol. 1, No. 2, 03.04.2015, p. 98-113.

Research output: Contribution to journal › Article › Academic › peer-review

TY - JOUR

T1 - A comparison of indices for identifying the number of clusters in hierarchical clustering

T2 - A study on cognition in schizophrenia patients

AU - Islam, Md Atiqul

AU - Alizadeh, Behrooz Z.

AU - van den Heuvel, Edwin R.

AU - Bruggeman, Richard

AU - Cahn, Wiepke

AU - de Haan, Lieuwe

AU - Kahn, René S.

AU - Meijer, Carin

AU - Myin-Germeys, Inez

AU - van Os, Jim

AU - Wiersma, Durk

PY - 2015/4/3

Y1 - 2015/4/3

N2 - Finding clusters in a complex dataset is not straightforward. Different indices were developed to quantify the number of clusters. Their performances were studied using unrealistic simulations, since they were considered at low dimensions. We investigated 14 indices for eight-dimensional data using simulations based on cognition measures. We focused on hierarchical clustering with Ward’s agglomerative technique. Results indicated that Duda and Hart, Hartigan and Gap/pc were best performing. They estimated the number of clusters within ±1 with high probabilities. Duda and Hart index was most consistent, while Gap/pc and WGap/pc together made a good distinction between single and multiple clusters.

AB - Finding clusters in a complex dataset is not straightforward. Different indices were developed to quantify the number of clusters. Their performances were studied using unrealistic simulations, since they were considered at low dimensions. We investigated 14 indices for eight-dimensional data using simulations based on cognition measures. We focused on hierarchical clustering with Ward’s agglomerative technique. Results indicated that Duda and Hart, Hartigan and Gap/pc were best performing. They estimated the number of clusters within ±1 with high probabilities. Duda and Hart index was most consistent, while Gap/pc and WGap/pc together made a good distinction between single and multiple clusters.

KW - Cluster analysis

KW - cluster indices

KW - hierarchical clustering

KW - homogeneous subgroups

KW - number of clusters

UR - http://www.scopus.com/inward/record.url?scp=85026827819&partnerID=8YFLogxK

U2 - 10.1080/23737484.2015.1103670

DO - 10.1080/23737484.2015.1103670

M3 - Article

AN - SCOPUS:85026827819

VL - 1

SP - 98

EP - 113

JO - Communications in Statistics Case Studies Data Analysis and Applications

JF - Communications in Statistics Case Studies Data Analysis and Applications

IS - 2

ER -

A comparison of indices for identifying the number of clusters in hierarchical clustering: A study on cognition in schizophrenia patients

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this