Weighted tests of homogeneity for testing the number of components in a mixture

Authors:
Edward Susko
Affiliations:
Department of Mathematics and Statistics, Dalhousie University, Halifax, NS, Canada
Venue:
Computational Statistics & Data Analysis
Year:
2003

Citing 0
Cited 3

Editorial: recent developments in mixture models

Computational Statistics & Data Analysis
Estimation for the number of components in a mixture model using stepwise split-and-merge EM algorithm

Pattern Recognition Letters
Testing the number of components of the mixture of two inverse Weibull distributions

International Journal of Computer Mathematics

Quantified Score

Hi-index	0.03

Visualization

Abstract

An important but difficult problem in fitting finite mixture models is estimating and testing the number of components in the mixture. Regularity conditions do not hold for large sample likelihood theory so that likelihood ratio tests cannot easily be implemented. However, a number of homogeneity tests have been developed to test for the presence of a mixture. Weighted versions of homogeneity tests are presented that can be used to test for the presence of additional components in a mixture. These tests are easily implemented, do not require long computational times and can incorporate covariates. Examples are given to illustrate the methodology and simulation results are presented that suggest that the tests have power comparable to the bootstrap likelihood ratio test.