R/ModelGAP.R

In citronmeliss/predictionpackr: Predictionpackr

#' Model GAP
#'
#' This function makes predictions with the GAP model.
#' @param data The study data frame. No default.
#' @export

ModelGAP <- function(data)

{
  ## Define model variables; bind gcs later
  model_variables <- c("age",
                       "sbp")

  ## Define cut points for binning
  cut_points <- list(age = c(0,60,Inf),
                     sbp = c(0,60,120, Inf))

  ## Define scores of variables
  scores <- list(age = c("3", "0"),
                 sbp = c("0", "4", "6"))

  ## Bin model variables with bin.model.variables
  binned_variables <- BinModelVariables(data,
                                        model_variables,
                                        cut_points,
                                        scores)

  ## Sum binned_variables to generate gap score. Then, invert.
  gap_predictions <- rowSums(cbind(binned_variables,
                                   data$gcs))

  return(gap_predictions)
}