make_constraint: Construct the constraint matrices for the spatial random...

In jtipton25/BayesMRA: Bayesian Multi-Resolution Gaussian Process Approximations

Construct the constraint matrices for the spatial random effects alpha

Description

Construct the constraint matrices for the spatial random effects alpha

Usage

make_constraint(MRA, constraint = "unconstrained", joint = TRUE)

Arguments

MRA	An MRA object. Usually contained in the output from mcmc_mra() or mcmc_mra_integrated() or for developers, the MRA object can be created directly from the function mra_wendland_2d()
constraint	What constraint should be applied to the spatial process? Options include no constraint (constraint = "unconstrained"), a constraint so the entire MRA process sums to 0 (constraint = "overall"), a constraint so that each of the M levels of the MRA process sum to 0 (constraint = "resolution"), or whether the predicte process must sum to 0 (constraint = "predicted"). Note: constraint = "predicted" is NOT currently supported.
joint	Should the spatial parameters alpha be sampled jointly or by each resolution

Value

A list of constraint matrices A_constraint and a_constraint for use in spam sampling functions

jtipton25/BayesMRA documentation built on Feb. 28, 2024, 1:27 p.m.