R/kern.R

In generalCorr: Generalized Correlations, Causal Paths and Portfolio Selection

#' Kernel regression with options for residuals and gradients.
#' 
#' Function to run kernel regression with options for residuals and gradients
#' asssuming no missing data.
#' 
#' @param dep.y {Data on the dependent (response) variable}
#' @param reg.x {Data on the regressor (stimulus) variables}
#' @param tol {Tolerance on the position of located minima of the cross-validation 
#' function (default =0.1)}
#' @param ftol {Fractional tolerance on the value of cross validation function
#'  evaluated at local minima (default =0.1)}
#' @param gradients {Make this TRUE if gradients computations are desired}
#' @param residuals {Make this TRUE if residuals are desired}
#' @importFrom np npreg npregbw
#' @return Creates a model object `mod' containing the entire kernel regression output.
#' Type \code{names(mod)} to reveal the variety of outputs produced by `npreg' of the `np' package.
#' The user can access all of them at will by using the dollar notation of R.
#' @note This is a work horse for causal identification.
#' @author Prof. H. D. Vinod, Economics Dept., Fordham University, NY
#' @seealso See \code{\link{kern_ctrl}}.
#' @references Vinod, H. D.'Generalized Correlation and Kernel Causality with 
#'  Applications in Development Economics' in Communications in 
#'  Statistics -Simulation and Computation, 2015, 
#'  \doi{10.1080/03610918.2015.1122048} 
#' @concept apd  amorphous partial derivative
#' @concept  kernel regression residuals
#' @concept  kernel regression gradients
#' @examples
#' 
#' \dontrun{
#' set.seed(34);x=matrix(sample(1:600)[1:50],ncol=2)
#' require(np); options(np.messages=FALSE)
#' k1=kern(x[,1],x[,2])
#' print(k1$R2) #prints the R square of the kernel regression
#' }
#' 
#' @export

kern <- function(dep.y, reg.x, tol = 0.1, ftol = 0.1, gradients = FALSE, residuals = FALSE) {
    gr = FALSE
    resz = FALSE
    if (gradients) 
        gr = TRUE
    if (residuals) 
        resz = TRUE
    # bandwidths for nonparametric regressions
    bw = npregbw(ydat = as.vector(dep.y), xdat = reg.x, tol = tol, ftol = ftol)
    mod = npreg(bws = bw, gradients = gr, residuals = resz)
    return(mod)
}