Kernel based Dirichlet sequences

DOI10.3150/22-BEJ1500MaRDI QIDQ6369071arXivFDO

Authors Patrizia Berti, Emanuela Dreassi, Fabrizio Leisen, Luca Pratelli, Pietro Rigo

Publication date 31 May 2021

Published in arXiv

Full work available at URL https://doi.org/10.3150/22-BEJ1500

Mathematics Subject Classification ID

Convergence of probability measures (60B10) Nonparametric estimation (62G05) Central limit and other weak theorems (60F05) Random measures (60G57) Exchangeability for stochastic processes (60G09)

arXiv classification

math.PR

Abstract: Let $X=(X_1,X_2,ldots)$ be a sequence of random variables with values in a standard space $(S,mathcalB)$. Suppose �egin{gather*} X_1sim

uquad ext{and}quad P�igl(X_{n+1}incdotmid X_1,ldots,X_n�igr)=frac{ heta

u(cdot)+sum_{i=1}^nK(X_i)(cdot)}{n+ heta}quadquad ext{a.s.} end{gather*} where $heta>0$ is a constant, $u$ a probability measure on $mathcalB$, and $K$ a random probability measure on $mathcalB$. Then, $X$ is exchangeable whenever $K$ is a regular conditional distribution for $u$ given any sub-$sigma$-field of $mathcalB$. Under this assumption, $X$ enjoys all the main properties of classical Dirichlet sequences, including Sethuraman's representation, conjugacy property, and convergence in total variation of predictive distributions. If $mu$ is the weak limit of the empirical measures, conditions for $mu$ to be a.s. discrete, or a.s. non-atomic, or $mullu$ a.s., are provided. Two CLT's are proved as well. The first deals with stable convergence while the second concerns total variation distance.

This page was built for publication: Kernel based Dirichlet sequences

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6369071)