Boot4PTCA: Compute the Bootstrapped factors scores for the I and J sets...

In HerveAbdi/PTCA4CATA: Partial Triadic Analysis (PTCA) for Check All That Apply (CATA) Data

Compute the Bootstrapped factors scores for the I and J sets from a Partial Triadic Correspondence analysis (PTCA).

Description

Boot4PTCA bootstraps the K-th dimension of a data cube and computes bootstrapped factor scores.

Usage

Boot4PTCA(
  ZeDataCube,
  fi,
  fj,
  eigs,
  nf2keep = 2,
  nBootIter = 100,
  compact = FALSE,
  eigen = FALSE,
  eigen.compact = TRUE
)

Arguments

ZeDataCube	An I * J * K data cube (K are observations) The third dimension (i.e., K) is bootstrapped.
fi	The factor scores for I (rows) from the epCA program
fj	The factor scores for J (columns) from the epCA program
eigs	The eigenvalues from the epCA program
nf2keep	how many factors to keep, default to 2
nBootIter	How many Bootstrap samples, default to 100 (RowsBoot = ZeBootCube_I, ColumnsBoot = ZeBootCube_J).
compact	(default = FALSE) if TRUE gives a compact version with only the results for the symmetric approach.
eigen	if FALSE compute also the bootstraped eigenvalues. NB It seems that the bootstrapped eigenvalues are biased (i.e., their mean is not equal to the corresponding eigenvalue). So this feature is experimental.
eigen.compact	when TRUE returns the whole matrix of bootstrapped eigenvalues. Default is FALSE. When eigen is FALSE, eigen.compact has no effect.

Value

With notation: I: number of rows (of ZeDataCube), J: number of columns (of ZeDataCube), L: number of factors kept (i.e., nf2keep), B: number of Bootstrap replicates (i.e., nBootIter); Boot4PTCA returns a list if compact FALSE: 1a) RowsBoot an I * L * B cube of Bootstrapped coordinates for the I-set 1b) RowsBoot.asym an I * L * B cube of Bootstrapped coordinates for the I-set (asymmetric projections); 2a) ColumnsBoot a J * L * B cube of Bootstrapped coordinates for the J-set if compact is FALSE 2b) ColumnsBoot.asym a J * L * B cube of Bootstrapped coordinates for the J-set.

Author(s)

Hervé Abdi

Examples

## Not run: 
BootFactorsIJ <- Boot4PTCAt(A.Cube.Of.Data,fi = fi, fj = fj, eigs = eigs)

## End(Not run)

HerveAbdi/PTCA4CATA documentation built on July 17, 2022, 5:41 a.m.