R/RcppExports.R

In smog: Structural Modeling by using Overlapped Group Penalty

# Generated by using Rcpp::compileAttributes() -> do not edit by hand
# Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393

#' Generalized linear model constraint on hierarchical structure
#' by using overlapped group penalty
#' 
#' @param y response variable, in the format of matrix. When family is 
#'          ``gaussian'' or ``binomial'', \code{y} ought to
#'          be a matrix of n by 1 for the observations; when family
#'          is ``coxph'', y represents the survival objects, that is,
#'          a matrix of n by 2, containing the survival time and the censoring status.
#'          See \code{\link[survival]{Surv}}.
#' @param x a model matrix of dimensions n by p,in which the column represents 
#'          the predictor variables.
#' @param g a numeric vector of group labels for the predictor variables.
#' @param v a numeric vector of binary values, represents whether or not the
#'          predictor variables are penalized. Note that 1 indicates
#'          penalization and 0 for not penalization.
#' @param lambda a numeric vector of three penalty parameters corresponding to L2 norm,
#'               squared L2 norm, and L1 norm, respectively.
#' @param hierarchy a factor value in levels 0, 1, 2, which represent different
#'                  hierarchical structure within groups, respectively. 
#'                  When \code{hierarchy=0}, \eqn{\lambda_2} and \eqn{\lambda_3} are 
#'                  forced to be zeroes; when \code{hierarchy=1}, \eqn{\lambda_2} is 
#'                  forced to be zero; when \code{hierarchy=2}, there is no constraint 
#'                  on \eqn{\lambda}'s. See \code{\link{smog}}.
#' @param family a description of the distribution family for the response
#'               variable variable. For continuous response variable,
#'               family is ``gaussian''; for multinomial or binary response
#'               variable, family is ``binomial''; for survival response
#'               variable, family is ``coxph'', respectively.
#' @param rho the penalty parameter used in the alternating direction method
#'            of multipliers algorithm (ADMM). Default is 10.
#' @param scale whether or not scale the design matrix. Default is \code{true}.
#' @param eabs  the absolute tolerance used in the ADMM algorithm. Default is 1e-3.
#' @param erel  the reletive tolerance used in the ADMM algorithm. Default is 1e-3.
#' @param LL    initial value for the Lipschitz continuous constant for
#'              approximation to the objective function in the Majorization-
#'              Minimization (MM) (or iterative shrinkage-thresholding algorithm
#'              (ISTA)). Default is 1.
#' @param eta gradient stepsize for the backtrack line search for the Lipschitz
#'            continuous constant. Default is 1.25.
#' @param maxitr the maximum iterations for convergence in the ADMM algorithm.
#'               Default is 500.
#'               
#' @seealso \code{\link{cv.smog}}, \code{\link{smog.default}}, \code{\link{smog.formula}}, 
#'          \code{\link{predict.smog}}, \code{\link{plot.smog}}.
#' 
#' @author Chong Ma, \email{chongma8903@@gmail.com}.
#' @references \insertRef{ma2019structural}{smog}
#' 
#' @return A list of 
#'         \item{coefficients}{A data frame of the variable name and the estimated coefficients}
#'         \item{llikelihood}{The log likelihood value based on the ultimate estimated coefficients}
#'         \item{loglike}{The sequence of log likelihood values since the algorithm starts}
#'         \item{PrimalError}{The sequence of primal errors in the ADMM algorithm}
#'         \item{DualError}{The sequence of dual errors in the ADMM algorithm}
#'         \item{converge}{The integer of the iteration when the convergence occurs}
#' 
#' 
#' @examples 
#' 
#' set.seed(2018) 
#' # generate design matrix x
#' n=50;p=100
#' s=10
#' x=matrix(0,n,1+2*p)
#' x[,1]=sample(c(0,1),n,replace = TRUE)
#' x[,seq(2,1+2*p,2)]=matrix(rnorm(n*p),n,p)
#' x[,seq(3,1+2*p,2)]=x[,seq(2,1+2*p,2)]*x[,1]
#' 
#' g=c(p+1,rep(1:p,rep(2,p)))  # groups 
#' v=c(0,rep(1,2*p))           # penalization status
#' 
#' # generate beta
#' beta=c(rnorm(13,0,2),rep(0,ncol(x)-13))
#' beta[c(2,4,7,9)]=0
#' 
#' # generate y
#' data1=x%*%beta
#' noise1=rnorm(n)
#' snr1=as.numeric(sqrt(var(data1)/(s*var(noise1))))
#' y1=data1+snr1*noise1
#' lambda = c(8,0,8)
#' hierarchy = 1
#' gfit1 = glog(y1,x,g,v,lambda,hierarchy,family = "gaussian")
#' 
#' @export
glog <- function(y, x, g, v, lambda, hierarchy, family = "gaussian", rho = 10, scale = TRUE, eabs = 1e-3, erel = 1e-3, LL = 1, eta = 1.25, maxitr = 1000L) {
    .Call(`_smog_glog`, y, x, g, v, lambda, hierarchy, family, rho, scale, eabs, erel, LL, eta, maxitr)
}

#' L1 proximal operator 
#' 
#' @param x numeric value.
#' @param lambda numeric value for the L1 penalty parameter.
#' 
#' @author Chong Ma, \email{chongma8903@@gmail.com}.
#' @references \insertRef{ma2019structural}{smog}
#' 
#' @return A numeric value soft-thresholded by \eqn{\lambda}, 
#'         which is \eqn{sign(x)(|x|-\lambda)_{+}}.
#' 
#' @examples 
#' proxL1(2.0,0.5)
#' 
#' @export
proxL1 <- function(x, lambda) {
    .Call(`_smog_proxL1`, x, lambda)
}

#' L2 proximal operator
#' 
#' @param x A vector of p numerical values.
#' @param lambda numeric value for the L2 penalty parameter.
#' 
#' @author Chong Ma, \email{chongma8903@@gmail.com}.
#' @references \insertRef{ma2019structural}{smog}
#' 
#' @return A numeric vector soft-thresholded by \eqn{\lambda} as a group, 
#'         which is \eqn{(1-\frac{\lambda \sqrt{p}}{\sqrt{x_1^2+\cdots+x_p^2}})_{+}\bm{x}}.
#' 
#' @examples
#' proxL2(rnorm(6,2,1),0.5)
#' 
#' @export
proxL2 <- function(x, lambda) {
    .Call(`_smog_proxL2`, x, lambda)
}

#' Composite proximal operator based on L2, L2-Square, and L1 penalties
#' 
#' @param x A numeric vector of two.
#' @param lambda a vector of three penalty parameters. \eqn{\lambda_1} and 
#'        \eqn{\lambda_2} are L2 and L2-Square (ridge) penalties for \eqn{x} in 
#'        a group level, and \eqn{\lambda_3} is the L1 penalty for \eqn{x_2}, respectively.
#' @param hierarchy a factor value in levels 0, 1, 2, which represent different
#'                  hierarchical structure in x, respectively. When \code{hierarchy=0},
#'                  \eqn{\lambda_2} and \eqn{\lambda_3} are forced to be zeroes; when
#'                  \code{hierarchy=1}, \eqn{\lambda_2} is forced to be zero; when 
#'                  \code{hierarchy=2}, there is no constraint on \eqn{\lambda}'s. 
#'                  See \code{\link{smog}}.
#' @param d indices for overlapped variables in x.   
#' 
#' @seealso \code{\link{cv.smog}}, \code{\link{smog.default}}, \code{\link{smog.formula}}, 
#'          \code{\link{predict.smog}}, \code{\link{plot.smog}}.
#' 
#' @author Chong Ma, \email{chongma8903@@gmail.com}.
#' @references \insertRef{ma2019structural}{smog}
#' 
#' @return A two-dimensional numerical vector, soft-thresholded based on a composition of 
#'         \eqn{\lambda_1}, \eqn{\lambda_2}, and \eqn{\lambda_3}.
#'         
#' @examples
#' prox(x = rnorm(6,2,1), lambda = c(0.5,0.3,0.1), hierarchy = 0, d = c(1,1,2,2,3,3))
#' 
#' @export
prox <- function(x, lambda, hierarchy, d) {
    .Call(`_smog_prox`, x, lambda, hierarchy, d)
}

#' Penalty function on the composite L2, L2-Square, and L1 penalties
#' 
#' @param x A vector of two numeric values, in which \eqn{x_1} represents
#'          the prognostic effect, and \eqn{x_2} for the predictive effect, 
#'          respectively. 
#' @param lambda a vector of three penalty parameters. \eqn{\lambda_1} and 
#'        \eqn{\lambda_2} are L2 and L2-Square (ridge) penalties for \eqn{x} in 
#'        a group level, and \eqn{\lambda_3} is the L1 penalty for \eqn{x_2}, respectively.
#' @param hierarchy a factor value in levels 0, 1, 2, which represent different
#'                  hierarchical structure in x, respectively. When \code{hierarchy=0},
#'                  \eqn{\lambda_2} and \eqn{\lambda_3} are forced to be zeroes; when
#'                  \code{hierarchy=1}, \eqn{\lambda_2} is forced to be zero; when 
#'                  \code{hierarchy=2}, there is no constraint on \eqn{\lambda}'s. 
#'                  See \code{\link{smog}}.
#' @param d indices for overlapped variables in x. 
#' 
#' @seealso \code{\link{cv.smog}}, \code{\link{smog.default}}, \code{\link{smog.formula}}, 
#'          \code{\link{predict.smog}}, \code{\link{plot.smog}}.
#' 
#' @author Chong Ma, \email{chongma8903@@gmail.com}.
#' @references \insertRef{ma2019structural}{smog}
#' 
#' @return A numeric value of the penalty function based on the composition of L2, L2-Square,
#'         and L2 penalties. 
#'         
#' @examples 
#' penalty(x = rnorm(6,2,1), lambda = c(0.5,0.3,0.1), hierarchy = 0, d = c(1,1,2,2,3,3))        
#' 
#' @export
penalty <- function(x, lambda, hierarchy, d) {
    .Call(`_smog_penalty`, x, lambda, hierarchy, d)
}

# Register entry points for exported C++ functions
methods::setLoadAction(function(ns) {
    .Call('_smog_RcppExport_registerCCallable', PACKAGE = 'smog')
})