fscca: Fast SCCA

# This file was generated by Rcpp::compileAttributes
# Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393

#' @export
split_in_groups <- function(length, k) {
    .Call('fscca_split_in_groups', PACKAGE = 'fscca', length, k)
}

#' @useDynLib fscca
#' @importFrom Rcpp evalCpp
NULL

#' NIPALS CCA algorithm
#'
#' @param X a matrix X that has been centered and scaled
#' @param Y a matrix Y that has been centered and scaled
#' @return a list containing a, b, u, v, and rho
#' @export
nipals <- function(X, Y) {
    .Call('fscca_nipals', PACKAGE = 'fscca', X, Y)
}

#' Sparse NIPALS CCA algorithm
#'
#' @param X a matrix X that has been centered and scaled
#' @param Y a matrix Y that has been centered and scaled
#' @param penalty_x A character string of type "lasso"
#' @param penalty_y A character string of type "lasso"
#' @param lamx a positive penalty on 'a'
#' @param lamy a positive penalty on 'b'
#' @return a list containing a, b, u, v, and rho (covariance)
#' @export
sparse_nipals <- function(X, Y, penalty_x, penalty_y, lamx, lamy) {
    .Call('fscca_sparse_nipals', PACKAGE = 'fscca', X, Y, penalty_x, penalty_y, lamx, lamy)
}

#' Fast sparse canonical correlation
#'
#' This is a fast implementation of sparse canonical correlation analysis.
#'
#' @param X a matrix of dimension n x p
#' @param Y a matrix of dimension n x q
#' @param penalty_x a string indicating the penalty function to use on matrix
#' X. Currently only "lasso" is implemented.
#' @param penalty_y a string indicating the penalty function to use on matrix
#' Y. Currently only "lasso" is implemented.
#' @param lam_x a numeric vector of tuning parameters on X
#' @param lam_y a numeric vector of tuning parameters on Y
#' @param k_folds an integer denoting the number of folds to use in cros-validation
#' @param n_components the number of components to compute
#' @param center center the columns to mean zero?
#' @param scale scale the columns to standard deviation one?
#' @return A list with matrices:
#' \describe{
#'     \item{A}{ A matrix of dimension p x n_components of the canonical
#' vector }
#'     \item{B}{ A matrix of dimension q x n_components of the canonical
#' vector }
#'     \item{U}{ A matrix of dimension n x n_components of the X * a }
#'     \item{V}{ A matrix of dimension n x n_components of the Y * b }
#'     \item{lambda}{ A matrix of dimension n_components x n of the optimal
#' tuning parameters}
#'     \item{covar}{ The average cross-validated covariance }
#' }
#' @export
fscca <- function(X, Y, penalty_x, penalty_y, lam_x, lam_y, k_folds = 5L, n_components = 1L, center = TRUE, scale = FALSE) {
    .Call('fscca_fscca', PACKAGE = 'fscca', X, Y, penalty_x, penalty_y, lam_x, lam_y, k_folds, n_components, center, scale)
}