|
Activity Number:
|
416
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #304203 |
|
Title:
|
A Separability Index for Clustering Problems and Its Applications
|
|
Author(s):
|
Arka Ghosh and Ranjan Maitra and Anna D. Peterson*+
|
|
Companies:
|
and Iowa State University and Iowa State University
|
|
Address:
|
126 Beedle Dr Apt 206, Ames, IA, 50014,
|
|
Keywords:
|
clustering ; gaussian ; algorithm ; index
|
|
Abstract:
|
We propose a separation index that captures the degree of difficulty in a clustering problem where each observation is generated from one of K different p-variate Gaussian distributions. This index is motivated by the intuition that an observation from a Gaussian distribution should be closer on average in distance to the mean of that distribution then to the mean of a different Gaussian distribution. The main purpose of this index is to be able to develop a data-simulation algorithm with a specified value of the index. Such data with varying values of the index can be used for comparison of various clustering algorithms. In contrast to similar algorithms in the literature, this algorithm does not impose any restriction on the model parameters (e.g. variance structures of the component distributions).
|
