Back to Annual Meeting Page		American Public Health Association 133rd Annual Meeting & Exposition December 10-14, 2005 Philadelphia, PA

Abstract #108752

Eric Harvey, PhD, Robert C. Lee, MS, and Patrick Crockett, PhD. Constella Health Sciences, Constella Group, 2605 Meridian Parkway, Durham, NC 27713, 919-313-7725, eharvey@constellagroup.com

The goal of cluster analysis is to identify homogenous subgroups within complex datasets. Many clustering techniques are available, some incorporating multivariate data. However, few of these techniques adequately handle multidimensional data. For example, suppose we want to cluster observations that have been measured in space and time. In order to cluster simultaneously on these four dimensional data, typical clustering methods first cluster on one dimension of the data and then re-cluster the results iteratively across the other dimensions. Such methods do not perform true multidimensional clustering since they do not consider all of the dimensions simultaneously. This type of data is becoming increasingly common with the widespread adoption of geographic information systems. We propose a true multidimensional clustering algorithm which can handle multivariate data. We demonstrate the method on simulated data and also apply it to a large dataset containing environmental pollutant measures and related health outcomes.

Learning Objectives:

• Understand the limitations of current clustering methods with regard to multidimensional data.
• Learn how to fit multidimensional mixture models to obtain cluster solutions.
• Visualize multidimensional cluster solutions.

Presenting author's disclosure statement:

I wish to disclose that I have NO financial interests or other relationship with the manufactures of commercial products, suppliers of commercial services or commercial supporters.