sample_lambda: Factor loading curve smoothing parameter sampler

In drkowal/FDLM: A Bayesian Multivariate Functional Dynamic Linear Model

Description

Sample the smoothing parameters for each factor loading curve.

Usage

sample_lambda(lambda, Psi, Omega = NULL, d = 1, uniformPrior = TRUE,
  orderLambdas = TRUE)

Arguments

lambda	K-dimensional vector of smoothing parameters (prior precisions) from previous MCMC iteration
Psi	J x K matrix of basis coefficients, where J is the number of basis functions and K is the number of factors
Omega	J x J penalty matrix; if NULL, assume it is diag(0, 0, 1,...,1)
d	dimension of tau; default is 1
uniformPrior	logical; when TRUE, use a uniform prior on prior standard deviations, 1/sqrt{lambda[k]}; otherwise use independent Gamma(0.001, 0.001) prior for each lambda[k]
orderLambdas	logical; when TRUE, enforce the ordering constraint lambda[1] > ... > lambda[K] for identifiability

Value

The K-dimensional vector of samoothing parameters, lambda.

drkowal/FDLM documentation built on May 20, 2019, 5:20 p.m.