|
Abstract:
|
This is a continuation of previous work on the spherical Dirichlet distribution and applications in text mining and gene expression. There is a need to model vectors at the positive orthant of the hypersphere, however, the previously proposed spherical Dirichlet distribution needs to use an appropriate transformation for handling singularities of 0 and 1. The proposed Delta-spherical Dirichlet distribution overcomes this difficulty for the case when the stochastic variables can take the values of 0 and 1. Basic properties of the proposed distribution, including normalizing constants and moments are developed. Estimators based on classical inferential statistics are obtained. An application using simulated data and a text mining example are developed and their results are discussed.
|
