Date: 24 Mar 2003

A Visual Method of Cluster Validation with Fastmap

* Final gross prices may vary according to local VAT.

Get Access

Abstract

This paper presents a visual method of cluster validation using the Fastmap algorithm. Two problems are tackled with Fastmap in the interactive process of discovering interesting clusters from real world databases. That is, (1) to verify separations of clusters created by a clustering algorithm and (2) to determine the number of clusters to be produced. They are achieved through projecting objects and clusters by Fastmap to the 2D space and visually examining the results by humans. We use a real example to show how this method has been used in discovering interesting clusters from a real data set.