Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering

John Nerbonne; Wilbert Heeringa; Franz Manni; P. Kleiweg

Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering

John Nerbonne
, Wilbert Heeringa
, Franz Manni
, P. Kleiweg

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

53 Citations (Scopus)

Abstract

Dialectometry produces aggregate DISTANCE MATRICES in which a distance is specified for each pair of sites. By projecting groups obtained by clustering onto geography one compares results with traditional dialectology, which produced maps partitioned into implicitly non-overlapping DIALECT AREAS. The importance of dialect areas has been challenged by proponents Of CONTINUA, but they too need to compare their findings to older literature, expressed in terms of areas.

Simple clustering is unstable, meaning that small differences in the input matrix can lead to large differences in results (Jain et al. 1999). This is illustrated with a 500-site data set from Bulgaria, where input matrices which correlate very highly (r = 0.97) still yield very different clusterings. Kleiweg et al. (2004) introduce COMPOSITE CLUSTERING, in which random noise is added to matrices during repeated clustering. The resulting borders are then projected onto the map.

The present contribution compares Kleiweg et al.'s procedure to resampled bootstrapping, and also shows how the same procedure used to project borders from composite clustering may be used to project borders from bootstrapping.

Original language	English
Title of host publication	Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007
Editors	C Preisach, H Burkhardt, L SchmidtThieme, R Decker
Place of Publication	BERLIN
Publisher	Springer
Pages	647-654
Number of pages	8
ISBN (Print)	978-3-540-78239-1
Publication status	Published - 2008
Event	31st Annual Conference of the German-Classification-Society - , Germany Duration: 7-Mar-2007 → 9-Mar-2007

Publication series

Name	STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION
Publisher	SPRINGER-VERLAG BERLIN
ISSN (Print)	1431-8814

Other

Other	31st Annual Conference of the German-Classification-Society
Country/Territory	Germany
Period	07/03/2007 → 09/03/2007

Handle.net

Persistent link

Cite this

Nerbonne, J., Heeringa, W., Manni, F., & Kleiweg, P. (2008). Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering. In C. Preisach, H. Burkhardt, L. SchmidtThieme, & R. Decker (Eds.), Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007 (pp. 647-654). (STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION). Springer.

Nerbonne, John ; Heeringa, Wilbert ; Manni, Franz et al. / Projecting dialect distances to geography : Bootstrap clustering vs. noisy clustering. Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007. editor / C Preisach ; H Burkhardt ; L SchmidtThieme ; R Decker. BERLIN : Springer, 2008. pp. 647-654 (STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION).

@inproceedings{6f492423a69a4db58a12a1c8718ecdbb,

title = "Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering",

abstract = "Dialectometry produces aggregate DISTANCE MATRICES in which a distance is specified for each pair of sites. By projecting groups obtained by clustering onto geography one compares results with traditional dialectology, which produced maps partitioned into implicitly non-overlapping DIALECT AREAS. The importance of dialect areas has been challenged by proponents Of CONTINUA, but they too need to compare their findings to older literature, expressed in terms of areas.Simple clustering is unstable, meaning that small differences in the input matrix can lead to large differences in results (Jain et al. 1999). This is illustrated with a 500-site data set from Bulgaria, where input matrices which correlate very highly (r = 0.97) still yield very different clusterings. Kleiweg et al. (2004) introduce COMPOSITE CLUSTERING, in which random noise is added to matrices during repeated clustering. The resulting borders are then projected onto the map.The present contribution compares Kleiweg et al.'s procedure to resampled bootstrapping, and also shows how the same procedure used to project borders from composite clustering may be used to project borders from bootstrapping.",

author = "John Nerbonne and Wilbert Heeringa and Franz Manni and P. Kleiweg",

note = "2008/j.nerbonne/pub004; 31st Annual Conference of the German-Classification-Society ; Conference date: 07-03-2007 Through 09-03-2007",

year = "2008",

language = "English",

isbn = "978-3-540-78239-1",

series = "STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION",

publisher = "Springer",

pages = "647--654",

editor = "C Preisach and H Burkhardt and L SchmidtThieme and R Decker",

booktitle = "Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft f{\"u}r Klassifikation e.V., Albert-Ludwigs Universit{\"a}t Freiburg, March 7-9, 2007",

}

Nerbonne, J, Heeringa, W, Manni, F & Kleiweg, P 2008, Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering. in C Preisach, H Burkhardt, L SchmidtThieme & R Decker (eds), Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007. STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION, Springer, BERLIN, pp. 647-654, 31st Annual Conference of the German-Classification-Society, Germany, 07/03/2007.

Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering. / Nerbonne, John; Heeringa, Wilbert; Manni, Franz et al.
Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007. ed. / C Preisach; H Burkhardt; L SchmidtThieme; R Decker. BERLIN: Springer, 2008. p. 647-654 (STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Projecting dialect distances to geography

T2 - 31st Annual Conference of the German-Classification-Society

AU - Nerbonne, John

AU - Heeringa, Wilbert

AU - Manni, Franz

AU - Kleiweg, P.

N1 - 2008/j.nerbonne/pub004

PY - 2008

Y1 - 2008

N2 - Dialectometry produces aggregate DISTANCE MATRICES in which a distance is specified for each pair of sites. By projecting groups obtained by clustering onto geography one compares results with traditional dialectology, which produced maps partitioned into implicitly non-overlapping DIALECT AREAS. The importance of dialect areas has been challenged by proponents Of CONTINUA, but they too need to compare their findings to older literature, expressed in terms of areas.Simple clustering is unstable, meaning that small differences in the input matrix can lead to large differences in results (Jain et al. 1999). This is illustrated with a 500-site data set from Bulgaria, where input matrices which correlate very highly (r = 0.97) still yield very different clusterings. Kleiweg et al. (2004) introduce COMPOSITE CLUSTERING, in which random noise is added to matrices during repeated clustering. The resulting borders are then projected onto the map.The present contribution compares Kleiweg et al.'s procedure to resampled bootstrapping, and also shows how the same procedure used to project borders from composite clustering may be used to project borders from bootstrapping.

AB - Dialectometry produces aggregate DISTANCE MATRICES in which a distance is specified for each pair of sites. By projecting groups obtained by clustering onto geography one compares results with traditional dialectology, which produced maps partitioned into implicitly non-overlapping DIALECT AREAS. The importance of dialect areas has been challenged by proponents Of CONTINUA, but they too need to compare their findings to older literature, expressed in terms of areas.Simple clustering is unstable, meaning that small differences in the input matrix can lead to large differences in results (Jain et al. 1999). This is illustrated with a 500-site data set from Bulgaria, where input matrices which correlate very highly (r = 0.97) still yield very different clusterings. Kleiweg et al. (2004) introduce COMPOSITE CLUSTERING, in which random noise is added to matrices during repeated clustering. The resulting borders are then projected onto the map.The present contribution compares Kleiweg et al.'s procedure to resampled bootstrapping, and also shows how the same procedure used to project borders from composite clustering may be used to project borders from bootstrapping.

M3 - Conference contribution

SN - 978-3-540-78239-1

T3 - STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION

SP - 647

EP - 654

BT - Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007

A2 - Preisach, C

A2 - Burkhardt, H

A2 - SchmidtThieme, L

A2 - Decker, R

PB - Springer

CY - BERLIN

Y2 - 7 March 2007 through 9 March 2007

ER -

Nerbonne J, Heeringa W, Manni F, Kleiweg P. Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering. In Preisach C, Burkhardt H, SchmidtThieme L, Decker R, editors, Data Analysis, Machine Learning and Applications. Proceedings of the 31st Annual Conference ofthe Gesellschaft für Klassifikation e.V., Albert-Ludwigs Universität Freiburg, March 7-9, 2007. BERLIN: Springer. 2008. p. 647-654. (STUDIES IN CLASSIFICATION, DATA ANALYSIS, AND KNOWLEDGE ORGANIZATION).

Projecting dialect distances to geography: Bootstrap clustering vs. noisy clustering

Abstract

Publication series

Other

Handle.net

Fingerprint

Cite this