Configurations and couplings: an exploratory study

Authors:
Warwick Graco;Hari Koesmarno
Affiliations:
Corporate Analytics, Australian Taxation Office, Canberra, Australia;Corporate Analytics, Australian Taxation Office, Canberra, Australia
Venue:
ICDM'13 Proceedings of the 13th international conference on Advances in Data Mining: applications and theoretical aspects
Year:
2013

Citing 6
Cited 0

Self-Organizing Maps

Self-Organizing Maps
Data Mining: Concepts and Techniques

Data Mining: Concepts and Techniques
Clustering high-dimensional data: A survey on subspace clustering, pattern-based clustering, and correlation clustering

ACM Transactions on Knowledge Discovery from Data (TKDD)
Data mining on multimedia data

Data mining on multimedia data
JMP Start Statistics: A Guide to Statistics and Data Analysis Using JMP, Fifth Edition

JMP Start Statistics: A Guide to Statistics and Data Analysis Using JMP, Fifth Edition
Coupled clustering ensemble: Incorporating coupling relationships both between base clusterings and objects

ICDE '13 Proceedings of the 2013 IEEE International Conference on Data Engineering (ICDE 2013)

Quantified Score

Hi-index	0.00

Visualization

Abstract

This is an exploratory study to see if configurations that were coupled to an output variable could be found in data. The focus in this study was on the modal configurations, which are profiles of best fit for clusters, and their average cluster scores for an output variable. A multistage procedure explained in the paper below was applied to a crime dataset to identify the modal configurations for a sample of cities and towns of the USA and their links to the incidence of violent crime. Three coupled configurations were found including one that was indicative of an African American Configuration having the highest rate of violent crime followed by one indicative of a High Divorce Configuration and one indicative of an Economic Hardship Configuration. The results indicated that using this multistage procedure is feasible for finding modal configurations and their couplings in data. The advantages of this approach are discussed and future directions with the research are outlined.