analyseBlocks: Analyse a list of data blocks containing shape information

In aharmer/morphoBlocks: Constructing a Multiple-Part Morphospace Using a Multiple-Block Method

Analyse a list of data blocks containing shape information

Description

Performs either Regularized Consensus Principal Component Analysis on a list of data blocks containing shape information, or performs principal component analysis on a superblock produced by the column-wise concatenation of the individual data blocks.

Usage

analyseBlocks(blockList, option = "rcpca", ncomp = 3)

Arguments

blockList	list of 'block' objects produced by the dodecBlock, formatBlock, readPts or readGPSA functions.
option	either option = "rcpca" (default) for Regularized Consensus Principal Component Analysis in mode 2 applied to the entire block list, or option = "pca" for principal component analysis applied to the superblock item in the block list.
ncomp	an integer specifying how many components should be calculated (default is 3)

Details

analyseBlocks is applied to an object of class "blockList" produced by the combineBlocks function and has two options: 1) option = "rcpca" and 2) option = "pca". The option = "rcpca" will perform Regularized Consensus Principal Component Analysis using the rgcca function from the RGCCA package (Tenenhaus and Guillemot 2017), and is the default option for analyseBlocks. The rgcca function itself has many options that each perform a different type of analysis. Here the analyseBlocks function is specifically calling the Regularized Consensus Principal Component Analysis in mode 2 option with scaling applied. For further detail see Tenenhaus and Guillemot (2017) and Tenenhaus et al. (2017). option = "pca" will perform principal component analysis on the superblock item in the block list using the prcomp function from base R.

Value

A list object containing output from the Regularized Consensus Principal Component Analysis or principal component analysis. The list contains the elements:

result	output from the Regularized Consensus Principal Component Analysis in mode 2 produced by the rgcca function from the RGCCA package, or output from principal component analysis produced by the prcomp function in base R.
option	either "rcpca" or "pca".
block.list	a list containing the data blocks and a concatenated superblock. Inherited from the supplied blockList object and retained for downstream analyses.
scores	component score values (for individual blocks and the consensus if option = "rpca"; for the superblock if option = "pca") (see scoresPlot for more detail).
block.loadings	component loadings (for individual blocks and the consensus if option = "rpca"; for the superblock if option = "pca") (see loadingsPlot for more detail).
p	number of points in the configurations of each data block. Inherited from the supplied blockList object and retained for downstream analyses.
k	number of dimensions that the points in each configuration has. Inherited from the supplied blockList object and retained for downstream analyses.
n	number of configurations included in each data block. Inherited from the supplied blockList object and retained for downstream analyses.

References

Tenenhaus A, Guillemot V. 2017. RGCCA: Regularized and Sparse Generalized Canonical Correlation Analysis for multiblock data 2.1.2. https://CRAN.R-project.org/package=RGCCA.

Tenenhaus M, Tenenhaus A, Groenen PJF. 2017. Regularized Generalized Canonical Correlation Analysis: A framework for sequential multiblock component methods. Psychometrika 82: 737-777 https://doi.org/10.1007/s11336-017-9573-x

Examples

block1 <- dodecBlock()
block2 <- dodecBlock()
blocklist <- combineBlocks(blocks = c(block1, block2))
result1 <- analyseBlocks(blocklist)
result2 <- analyseBlocks(blocklist, option = "pca", ncomp = 10)

aharmer/morphoBlocks documentation built on Nov. 21, 2022, 5:45 p.m.