Abstract:We propose a Bayesian nonparametric clustering approach to study the spatial heterogeneity effect for functional data observed at spatially correlated locations. We consider a geographically weighted Chinese restaurant process equipped with a conditional autoregressive prior to capture fully the spatial correlation of function curves. To sample efficiently from our model, we customize a prior called Quadratic Gamma, which ensures conjugacy. We design a Markov chain Monte Carlo algorithm to infer simultaneously the posterior distributions of the number of groups and the grouping configurations. The superior numerical performance of the proposed method over competing methods is demonstrated using simulated examples and a U.S. annual precipitation study. Résumé Dans ce travail, les auteurs présentent une approche de regroupement non paramétrique bayésienne pour examiner l'hétérogénéité spatiale dans des données fonctionnelles observées à des sites corrélés. Leur méthode s'appuie sur un processus du restaurant chinois pondéré géographiquement, complété par un a priori autorégressif conditionnel, pour capturer la corrélation spatiale des courbes fonctionnelles. Afin d'échantillonner efficacement dans ce cadre, ils introduisent un a priori Gamma quadratique qui garantit la propriété de conjugaison. Ils proposent également un algorithme de Monte‐Carlo par chaînes de Markov pour estimer les lois a posteriori du nombre de groupes et de leurs configurations. À travers des simulations et une étude sur les précipitations annuelles aux États‐Unis, ils démontrent que la performance de leur méthode est souvent supérieure à celle d'approches alternatives.

Estimating the Number of Clusters Via Proportional Chinese Restaurant Process

Reducing over-clustering via the powered Chinese restaurant process

Powered Dirichlet Process for Controlling the Importance of "Rich-Get-Richer" Prior Assumptions in Bayesian Clustering

A Non-MCMC Procedure for Fitting Dirichlet Process Mixture Models

Dirichlet Process Parsimonious Mixtures for clustering

A Clustering Method Based on Dirichlet Process Mixture Model

Distance Dependent Chinese Restaurant Processes

Robust modeling of mixture probabilistic principal component analysis and process monitoring application

CDP Mixture Models for Data Clustering

Short Text Clustering Based on Pitman-Yor Process Mixture Model

A Bayesian approach to disease clustering using restricted Chinese restaurant processes

A simple proof of Pitman-Yor's Chinese restaurant process from its stick-breaking representation

Clustering spatial functional data using a geographically weighted Dirichlet process

Non-exchangeable random partition models for microclustering

Identification of piecewise affine model for batch processes based on constrained clustering technique

DP-Dip: A Skinny Method for Estimating the Number and Center of Clusters

Concentration in the Generalized Chinese Restaurant Process

Multimode process data modeling: A Dirichlet process mixture model based Bayesian robust factor analyzer approach

A Comparative Study of Novel Robust Clustering Algorithms

Robust Dirichlet Process mixtures.

A Tutorial on Dirichlet Process Mixture Modeling.