bdCCA_hdf5: Canonical Correlation Analysis

In BigDataStatMeth: Statistical Methods and Algorithms for Big Data

Canonical Correlation Analysis

Description

This function is an application of the BigDataStatMeth functions to generate new methods. This function perform a Canonical Correlation Analysis from two matrices stored in hdf5 data file. This function applies matrix partitioning, merge bloks to create a full matrix, apply a function to different blocks, etc.

Usage

bdCCA_hdf5(
  filename,
  X,
  Y,
  m = 10,
  bcenter = TRUE,
  bscale = FALSE,
  bycols = FALSE,
  overwriteResults = FALSE,
  keepInteResults = FALSE,
  threads = 1,
  k = 4,
  q = 1
)

Arguments

filename	string file name where dataset to normalize is stored.
X	Dataset, path inside the hdf5 data file.
Y	Dataset, path inside the hdf5 data file.
m	Integer, number of blocks in which we want to partition the matrix to perform the calculations.
bcenter,	Boolean, if true, dataset is centered to perform calculus.
bscale,	Boolean, if true, dataset is centered to perform calculus.
bycols,	Boolean by default = true, true indicates that the imputation will be done by columns, otherwise, the imputation will be done by rows.
overwriteResults,	Boolean, if true, datasets existing inside a file must be overwritten if we are using the same names.
keepInteResults,	Boolean, if false, intermediate results will be removed.
threads	(optional) only used in some operations inside function. If threads is null then threads = maximum number of threads available - 1.
k	(optional) number of local SVDs to concatenate at each level
q	(optional) number of levels

Value

hdf5 data file with CCA results,

Examples

   print ("Example in vignette")

BigDataStatMeth documentation built on March 30, 2022, 1:07 a.m.