pCCA: Regularaized Sparse Canonical Covariates Analysis

In mortenarendt/StructuralKnowledgeModl: Incorporation of Structural Knowledge in Multivariate Models

Description

This functions calculates a CCA model with L1 norm sparse penalties on the canonical loadings, as well as L2 norm penalties on the structural loadings in Y.

Usage

pCCA(X, Y, R = diag(1, ncol(Y)), ncomp = 1, sumabs = c(sqrt(ncol(X)),
  sqrt(ncol(Y))), na = 30, nperm = 100)

Arguments

X	a n by px matrix
Y	a n by py matrix
R	a py by py symmetric matrix indicating strucural relationship between features in Y.
ncomp	= 1 (default) - number of components
sumabs	= sqrt(c(px,py)) - vector of L1 contraint on the canonical loadings u and v respectively
na	= 30 (default) - indicating the grid of structural penalties based on R to tests or a vector of specific penalties.
nperm	= 100 (default) - number of permutations to be perfomed

mortenarendt/StructuralKnowledgeModl documentation built on May 21, 2019, 11:42 a.m.