lambdaseq_generate {mixedCCA}

Internal data-driven lambda sequence generating function.

Description

This internal function generates lambda sequence of length nlamseq equally spaced on a logarithmic scale. Since this is for sparse CCA, it returns a list of two vectors. Each vector will be used for each data set X1 and X2. And w1 and w2 denote canonical vector for each data set.

Usage

lambdaseq_generate(
  nlamseq = 20,
  lam.eps = 0.01,
  Sigma1,
  Sigma2,
  Sigma12,
  w1init = NULL,
  w2init = NULL
)

Arguments

nlamseq

The length of lambda sequence

lam.eps

The smallest value for lambda as a fraction of maximum lambda value

Sigma1

Covariance/correlation matrix of X1 (p1 by p1)

Sigma2

Covariance/correlation matrix of X2 (p2 by p2)

Sigma12

Covariance/correlation matrix between X1 and X2

w1init

Initial value for canonical vector w1

w2init

Initial value for canonical vector w2

Value

lambdaseq_generate returns a list of length 2. Each vector is of the same length nlamseq and will be used for each data set separately.