R/report_MSE.R

In CorReg: Linear Regression Based on Linear Structure Between Variables

#' Quickly reports some MSE
#' 
#' @description Some MSE reporting to have an overview of the predictive quality of a model.
#' @param real a numeric vector that contains the true values
#' @param prediction a numeric vector that contains the predicted values.
#' 
#' @return a list containing the MSE values. 
#' RMSE is the root MSE (square root of the mean square error)
#' Relative is the mean of the relative errors (Root errors divided by the real values)
#' !standard_Deviation is the standard deviation of the root errors
#' 
#' @export
report_MSE <- function(real, prediction){
   RMSE = mean(sqrt((prediction-real)^2))
   RelRMSE = mean(sqrt((prediction-real)^2)/real)*100
   SDRSE = sd(sqrt((prediction-real)^2))
   MAE = mean(abs(prediction-real)) # Mean Absolute Error
   relMAE = mean(abs(prediction-real)/real)*100 # Mean Absolute Error
   
   return(list(RMSE = RMSE, RelRMSE = RelRMSE, MAE = MAE, relMAE = relMAE, Standart_Deviation = SDRSE))
}