init_bspbss {BSPBSS}

Initial values

Description

Generate initial values, set up priors and perform kernel decomposition for the MCMC algorithm.

Usage

init_bspbss(
  X,
  coords,
  rescale = TRUE,
  center = FALSE,
  q = 2,
  dens = 0.5,
  ker_par = c(0.05, 20),
  num_eigen = 500,
  noise = 0
)

Arguments

X

Data matrix with n rows (sample) and p columns (voxel).

coords

Cordinate matrix with p rows (voxel) and d columns (dimension).

rescale

If TRUE, rows of X are rescaled to have unit variance.

center

If TRUE, rows of X are mean-centered.

q

Number of latent sources.

dens

The initial density level (between 0 and 1) of the latent sources.

ker_par

2-dimensional vector (a,b) with a>0, b>0, specifing the parameters in the modified exponetial squared kernel.

num_eigen

Number of eigen functions.

noise

Gaussian noise added to the initial latent sources, with mean 0 and standard deviation being noise * sd(S0), where sd(S0) is the standard deviation of the initial latent sources.

Value

List containing initial values, priors and eigen functions/eigen values of the kernel of the Gaussian process.

Examples

sim = sim_2Dimage(length = 30, sigma = 5e-4, n = 30, smooth = 6)
ini = init_bspbss(sim$X, sim$coords, q = 3, ker_par = c(0.1,50), num_eigen = 50)