vbSearchControl: Parameters for variable block structure search.

In HDclust: Clustering High Dimensional Data with Hidden Markov Model on Variable Blocks

Parameters for variable block structure search.

Description

This function creates a list with parameters for the search of a variable block structure used as an argument for hmmvbTrain.

Usage

vbSearchControl(
  perm = NULL,
  numstPerDim = NULL,
  dim = NULL,
  maxDim = 10,
  minDim = 1,
  nperm = 1,
  relax = FALSE
)

Arguments

perm	A list of integer vectors specifying variable permutations. If provided, the argument dim must be supplied.
numstPerDim	An integer vector of length maxDim specifying a map from the variable block dimensionality to the number of states in the block. kth value in the vector corresponds to number of states for dimensionality k.
dim	Data dimensionality. Must be provided with perm argument, otherwise is ignored.
maxDim	Maximum variable block dimension.
minDim	Minimum variable block dimension. Should be an integer equal to 1 or 2.
nperm	The number of variable permutations. This parameter is ignored if permutations are provided in perm argument.
relax	A logical value indicating whether or not variable block structure search will be performed under less restricting conditions.

Value

The named list with parameters.

See Also

hmmvbTrain

Examples

# setting up permutations
perm <- list(c(1,2,3), c(1,3,2), c(3,2,1))
searchControl <- vbSearchControl(perm=perm, dim=3)

# setting up a map between block dimensionality and number of states
searchControl <- vbSearchControl(maxDim=5, numstPerDim=c(3,4,5,6,7))

HDclust documentation built on Sept. 20, 2024, 5:09 p.m.