gpdpPost: Run a Bayesian functional data model under a GP prior whose...

In growfunctions: Bayesian Non-Parametric Dependent Models for Time-Indexed Functional Data

Run a Bayesian functional data model under a GP prior whose parameters employ a DP prior

Description

An internal function to gpdpgrow

Usage

gpdpPost(
  y,
  ipr,
  Omega_t,
  Omega_s,
  gp_mod,
  jitter,
  gp_shape,
  gp_rate,
  noise_shape,
  noise_rate,
  lower,
  upper,
  w_star,
  w,
  n_slice_iter,
  y_index,
  n.iter,
  n.burn,
  n.thin,
  n.tune,
  M_init,
  dp_shape,
  dp_rate,
  progress
)

Arguments

y	An N x T matrix of N observations of T x 1 functions
ipr	An optional input vector of inclusion probabilities for each observation unit in the case the observed data were acquired through an informative sampling design, so that unbiased inference about the population requires adjustments to the observed sample. Defaults to ipr = rep(1,nrow(y)) indicating an iid sample.
Omega_t	A T x T matrix of squared Eucidean distances for T time points
Omega_s	A list object of length L_s, where each contains the T x T matrix of Euclidean distances associated to each seasonal covariance term.
gp_mod	An L x 1 numeric vector denoting the selected covariance function for each of L terms. gp_mod = 1 is "rq". gp_mod = 2 is "se". gp_mod = 3 is "sn".
jitter	Numeric value added to diagonals of GP covariance matrix to stabilize inversion.
gp_shape	The shape parameter of the Gamma base distribution for the kappa_star locations used to sample the DP prior on the P GP covariance parameters, kappa, for each experimental unit.
gp_rate	The rate parameter of the Gamma base distribution for the kappa_star locations used to sample the DP prior on the P GP covariance parameters, kappa, for each experimental unit.
noise_shape	The shape parameter of the Gamma base distribution on tau_e, the model noise precision parameter. Defaults to noise_shape = 3.
noise_rate	The rate parameter of the Gamma base distribution on tau_e, the model noise precision parameter. Defaults to noise_rate = 1.
lower	Minimum in range of support for GP covariance parameters, kappa.
upper	Maximum in range of support for GP covariance parameters, kappa.
w_star	Tuning parameter for number of locations to sample not linked to observations in the auxiliary Gibbs sampler for cluster assignments.
w	Tuning parameter for slice sampling interval width used for GP covariance parameters, kappa.
n_slice_iter	Maximum number of steps to widen slice samplind width for GP covariance parameters, kappa.
y_index	List object where each contains index of time points to use in n progressively coarser distribution for sampling kappa in tempered update steps.
n.iter	The number of MCMC sampling iterations
n.burn	The number of warm-up iterations to discard
n.thin	The interval or step size of post-burn-in samples to return
n.tune	The number of tuning iterations to update the slice sampler width, w.
M_init	Starting value of number of clusters for sampling cluster assignments.
dp_shape	The shape parameter for the \Gamma prior on the DP concentration parameter. The rate parameter is set of 1.
dp_rate	The rate parameter for the \Gamma prior on the DP concentration parameter. Default value is 1.
progress	An indicator in {0,1} denoting whether to display a progress bar during model execution. progress = 1 displays a progress bar. Defaults to progress = 1.

Value

res A list object containing MCMC runs for all model parameters.

Note

Intended as an internal function for gpdpgrow

Author(s)

Terrance Savitsky [e-mail]

See Also

gpdpgrow

growfunctions documentation built on June 22, 2024, 11:49 a.m.