Efficient likelihood computations for some multivariate Gaussian Markov random fields

JOURNAL OF MULTIVARIATE ANALYSIS

卷:168

Article

Ippoliti, L.¹  Martin, R. J.¹  Romagnoli, L.² 

[1] Univ G dAnnunzio, Dipartimento Econ, Viale Pindaro 42, I-65127 Pescara, Italy

[2] Univ Molise, Dipartimento Econ, Via F De Sanctis, I-86100 Campobasso, Italy

关键词: Conditional autoregressive model;    Gaussian Markov random fields;    Lattice data;    Maximum likelihood estimation;    Multivariate observations;    Regional data;   

DOI  :  10.1016/j.jmva.2018.07.007

来源: Elsevier

【 摘 要 】

Data collected from spatial locations are often multivariate. Gaussian conditional autore-gressive (CAR) models, also known as Gaussian Markov random fields, are frequently used to analyze such continuous data, or as models for the parameters of discrete distributions. Two difficulties in Gaussian maximum likelihood estimation are ensuring that the parameter estimates are allowable values, and computing the likelihood efficiently. It is shown here that, for some commonly-used multivariate CAR models, checking for allowable parameter values can be facilitated, and the likelihood can be computed very quickly. (C) 2018 Elsevier Inc. All rights reserved.

【 授权许可】

Free