Novel Mixtures Based on the Dirichlet Distribution: Application to Data and Image Classification
The Dirichlet distribution offers high flexibility for modeling data. This paper describes two new mixtures based on this density: the GDD (Generalized Dirichlet Distribution) and the MDD (Multinomial Dirichlet Distribution) mixtures. These mixtures will be used to model continuous and discrete data, respectively. We propose a method for estimating the parameters of these mixtures. The performance of our method is tested by contextual evaluations. In these evaluations we compare the performance of Gaussian and GDD mixtures in the classification of several pattern-recognition data sets and we apply the MDD mixture to the problem of summarizing image databases.
KeywordsExpectation Maximization Algorithm Dirichlet Distribution Initialization Method Natural Gradient Wisconsin Breast Cancer
Unable to display preview. Download preview PDF.
- 2.Bouguila, N., Ziou, D. and Vaillancourt, J. The Introduction of Dirichlet Mixture into Image Processing Applications. Submitted to. IEEE Transactions on Image Processing.Google Scholar
- 9.Kaufman, L. and Rousseeuw, P.J. Finding Groups in Data. John Wiley, New York, 1990.Google Scholar
- 10.Kherfi, M.L., Ziou, D. and Bernardi, A. Content-Based Image Retrieval Using Positive and Negative Examples. To appear, 2002.Google Scholar
- 13.Rao, C.R. Advanced Statistical Methods in Biomedical Research. New York: John Wiley and Sons, 1952.Google Scholar