This work describes a Bayesian approach for model selection in Gaussian conditional autoregressive models and Gaussian simultaneous autoregressive models which are commonly used to describe spatial lattice data. The approach is aimed at situations when all competing models have the same mean structure, but differ on some aspects of their covariance structures. The proposed approach uses as sele...

We present a methodology for generating probabilistic predictions for the Disturbance Storm Time (Dst) geomagnetic activity index. We focus on the One Step Ahead prediction task and use the OMNI hourly resolution data to build our models. Our proposed methodology is based on the technique of Gaussian Process Regression. Within this framework we develop two models; Gaussian Process Autoregressiv...

Parameters of Gaussian spatial models are often estimated using the maximum likelihood approach. In spite of its merits, this methodology is not practical for large data sets. We study the asymptotic properties of the estimators that minimize three alternatives to the likelihood function, which are meant to increase the computational efficiency. This is achieved by applying the information sand...

Parameters of Gaussian multivariate models are often estimated using the maximum likelihood approach. In spite of its merits, this methodology is not practical when the sample size is very large, as, for example, in the case of massive georeferenced data sets. In this paper, we study the asymptotic properties of the estimators that minimize three alternatives to the likelihood function, designe...

1. By using the sampling theorem to expand N (t − τ, ω) and h(τ) with respect to τ we obtain: Z(t, ω) = τ k∈Z h(kT) sinc π τ − kT T k∈Z N (t − kT, ω) sinc π τ − kT T dτ = T k∈Z h(kT)N (t − kT, ω) where we used orthogonality of the set of functions {sinc (π(τ − kT)/T)} k∈Z. 2. Since N (τ) is a complex Gaussian random process, for each finite set of epochs t 1 , ..., t K , the random variables {N...