R/F_getRowMat.R

In CenterForStatistics-UGent/RCM: Fit row-column association models with the negative binomial distribution for the microbiome

#' Return a matrix of row scores
#'
#' @param sampleScore a vector of length n with sample scores
#' @param responseFun a character string, the type of response function,
#'  either 'linear' or 'quadratic'
#' @param NB_params a v-by-p matrix of parameters of theresponse function
#' @param taxonCoef A vector of coefficients
#' @param spline The cubic smoothing spline
#'
#' Multiplying the old offset with the exponent matrix times
#' the importance parameter obtains the new one based on lower dimension
#'
#' @return a n-by-p matrix of scores
#' @importFrom VGAM predict
getRowMat = function(sampleScore, responseFun, 
    NB_params, taxonCoef, spline) {
    if (responseFun == "nonparametric") {
        cbind(1, sampleScore, predict(spline, 
            x = sampleScore)$y) %*% c(taxonCoef, 
            1)
    } else {
        buildDesign(sampleScore, responseFun) %*% 
            NB_params
    }
}