R/mark.R

In jr-packages/jrPred: Predictive Analytics 1 day by Jumping Rivers

#' A function for marking models during the predictive analytics course
#' 
#' @param model A model resulting from a call to train
#' @return NULL
#' @importFrom stats lowess
#' @export
mark = function(model){
  if(!inherits(model,"validated")) stop("Make sure to validate your model before trying to mark it")
  env = new.env()
  data(FuelEconomy, package = "AppliedPredictiveModeling", envir = env)
  cars2011 = env$cars2011
  pred = predict(model, cars2011)
  err = cars2011$FE - pred
  rmse = sqrt(mean(err*err))
  col1 = rgb(204,74,83, maxColorValue=255)
  
  op = par(mfrow = c(1,2), oma = c(0,0,1,0),
           mar=c(3,3,2,1), 
           mgp=c(2,0.4,0), tck=-.01,
           cex.axis=0.9, las=1)
  on.exit(par(op))
  plot(cars2011$FE,pred,xlab = "Observed", ylab = "Fitted", col = "black", 
       xlim = range(cars2011$FE),
       ylim = range(cars2011$FE), 
       pch=21, bg="grey90", panel.first=grid())
  abline(0,1)
  lines(lowess(cars2011$FE,pred),col = col1, lwd = 2.5, lty = 2)
  plot(cars2011$FE, err, xlab = "Observed", ylab = "Errors", col = "black", 
       pch=21, bg="grey90", panel.first=grid())
  abline(h = 0)
  lines(lowess(cars2011$FE,err),col = col1, lwd = 2, lty = 2)
  mtext(text = paste("RMSE: ",round(rmse,4)), outer = TRUE, line = -1)
}