Posterior convergence rate of a class of Dirichlet process mixture model for compositional data

详细信息    查看全文

• 作者：André ; s F. Barrientos^a ; ^{afb26@stat.duke.edu" class="auth_mail" title="E-mail the corresponding author} ; Alejandro Jara^b ; ^{atjara@uc.cl" class="auth_mail" title="E-mail the corresponding author} ; Claudia Wehrhahn^c ; ^{claudia@sigma.iimas.unam.mx" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Simplex ; Bayesian nonparametrics ; Frequentist asymptotics
• 刊名：Statistics & Probability Letters
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：120
• 期：Complete
• 页码：45-51
• 全文大小：442 K

文摘

We propose a Dirichlet process mixture of mixtures of Dirichlet models for density estimation. By assuming random sampling from a density belonging to a Hölder class, we show that the posterior distribution of the model is rate-optimal.