Associating Families of Curves Using Feature Extraction and Cluster Analysis

Jane L. Terry, Andrew Crampton, Chris J. Talbot

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


The focus of this paper is to provide a reliable approach for associating families of curves from within a large number of curves. The method developed assumes that it is not known how many families are present, or how many curves are held within a family. The algorithm described has been developed for use on acoustical data, where there is a strong physical relationship between related curves. In the solution to the problem each of the curves have several key features which are measured and parametrised. This results in the characteristics of each curve being described by a small number of directly comparable parameters. Using these parameters it is then possible to find the related curves by applying cluster analysis
Original languageEnglish
Title of host publicationAlgorithms for Approximation
Subtitle of host publicationProceedings of the 5th International Conference, Chester, July 2005
EditorsArmin Iske, Jeremy Levesley
Place of PublicationBerlin
Number of pages10
ISBN (Electronic)9783540465515
ISBN (Print)9783540332831, 9783642069949
Publication statusPublished - 2007
Event5th International Conference on Algorithms for Approximation - Chester, United Kingdom
Duration: 17 Jul 200521 Jul 2005
Conference number: 5


Conference5th International Conference on Algorithms for Approximation
Country/TerritoryUnited Kingdom


Dive into the research topics of 'Associating Families of Curves Using Feature Extraction and Cluster Analysis'. Together they form a unique fingerprint.

Cite this