WIT Press

Evaluation Of Hierarchical Clustering Algorithms As Classifiers Of Hypertensive Blood Pressure And Heart Rate Recordings

Price

Free (open access)

Paper DOI

10.2495/BIO970231

Volume

4

Pages

10

Published

1997

Size

964 kb

Author(s)

A.N. Kastania & M.P. Bekakos

Abstract

One of the biggest problems in hierarchical cluster analysis is validity of the classification results since various methods are available. This study evaluates hierarchical clustering algorithms as classifiers of 24h hypertensive blood pressure and heart rate recordings based on an algorithm designed and developed for this purpose. It also attempts to provide a comparison of the derived clusters with the results of a pattern recognition algorithm for biomedical waveform analysis in order to clarify the limitations on the use of hierarchical clustering methods for the classification of biomedical waveforms. 1 Introduction Cluster analysis is a tool of exploratory data analysis that

Keywords