Main function for DIVAS analysis. Given a list of data blocks with matched columns (samples), will return identified joint structure with diagnostic plots.
Usage
DIVASmain(
datablock,
nsim = 400,
iprint = TRUE,
colCent = FALSE,
rowCent = FALSE,
figdir = NULL
)Arguments
- datablock
A list of matrices with the same number of columns (samples).
- nsim
Number of bootstrap resamples for inferring angle bounds.
- iprint
Whether to print diagnostic figures.
- colCent
Whether to column centre the input data blocks.
- rowCent
Whether to row centre the input data blocks.
- figdir
If not NULL, then diagnostic plots will be saved to this directory.
Value
A list containing DIVAS integration results. Most important ones include
- matBlocks
List of scores representing shared and partially shared joint structures.
- matLoadings
List of loadings linking features in each data block with scores.
- keyIdxMap
Mapping between indices of the previous lists and data blocks.
See Details for more explanations.
Details
DIVASmain returns a list containing all important information returned from the DIVAS algorithm. For users, the most important ones are scores (matBlocks), loadings (matLoadings) and an index book (keyIdxMap) explaining what joint structures each score or loading matrix correpsond to.
matBlocks is a list containing scores. Each element of matBlocks is indexed by a number. For example, suppose one of the indices is "7", then keyIdxMap[["7"]] contains indices of data blocks corresponding to the index 7. That is, matBlocks[["7"]] contains the scores for all samples representing the joint structures of data blocks in keyIdxMap[["7"]].