R/parameterpower.R

In skpr: Design of Experiments Suite: Generate and Evaluate Optimal Designs

#'@title Calculates parameter power
#'
#'@description Calculates parameter power
#'
#'@param RunMatrix The run matrix
#'@param levelvector The number of levels in each parameter (1st is always the intercept)
#'@param anticoef The anticipated coefficients
#'@param alpha the specified type-I error
#'@param vinv The V inverse matrix
#'@return The parameter power for the parameters
#'@keywords internal
parameterpower = function(RunMatrix, levelvector=NULL, anticoef, alpha, vinv = NULL, degrees=NULL, parameter_names) {
  #Generating the parameter isolating vectors
  q = vector("list", dim(attr(RunMatrix, "modelmatrix"))[2])
  for (i in 1:length(q)) {
    vec = rep(0, dim(attr(RunMatrix, "modelmatrix"))[2])
    vec[i] = 1
    q[[i]] = t(vec)
  }
  if (is.null(degrees)) {
    degrees = rep(dim(attr(RunMatrix, "modelmatrix"))[1] -
                    dim(attr(RunMatrix, "modelmatrix"))[2],length(q))
  } else {
    degrees_long = rep(0,length(q))
    for (i in 1:length(parameter_names)) {
      for (term in parameter_names[[i]]) {
        degrees_long[which(term == colnames(attr(RunMatrix, "modelmatrix")))] = degrees[i]
      }
    }
    degrees_long[is.na(degrees_long)] = dim(attr(RunMatrix, "modelmatrix"))[1]-dim(attr(RunMatrix, "modelmatrix"))[2]
    degrees = degrees_long
  }

  power = c(length(q))
  for (j in 1:length(q)) {
    if(degrees[j] != 0) {
      power[j] = calculatepower(attr(RunMatrix, "modelmatrix"), q[[j]],
                                calcnoncentralparam(attr(RunMatrix, "modelmatrix"),
                                                    q[[j]], anticoef, vinv = vinv),
                                alpha, degrees[j])
    } else {
      power[j] = NA
    }
  }
  return(power)
}