Two Dimensional Projection Pursuit Applied to Gaussian Mixture Model Fitting
In this paper we seek a Gaussian mixture model (GMM) of an n-variate probability density function. Usually the parameters of GMMs are determined by a maximum likelihood (ML) criterion. A practical deficiency of ML fitting of GMMs is poor performance when dealing with high-dimensional data since a la...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
International Institute of Informatics and Cybernetics
2003-08-01
|
| Series: | Journal of Systemics, Cybernetics and Informatics |
| Subjects: | |
| Online Access: | http://www.iiisci.org/Journal/CV$/sci/pdfs/P689419.pdf
|
| Summary: | In this paper we seek a Gaussian mixture model (GMM) of an n-variate probability density function. Usually the parameters of GMMs are determined by a maximum likelihood (ML) criterion. A practical deficiency of ML fitting of GMMs is poor performance when dealing with high-dimensional data since a large sample size is needed to match the accuracy that is possible in low dimensions. We propose a method to fit the GMM to multivariate data which is based on the two-dimensional projection pursuit (PP) method. By means of simulations we compare the proposed method with a one-dimensional PP method for GMM. We conclude that a combination of one- and twodimensional PP methods could be useful in some applications. |
|---|---|
| ISSN: | 1690-4524 |