MAE: Mean Absolute Error

In DiceEval: Construction and Evaluation of Metamodels

Mean Absolute Error

Description

The mean of absolute errors between real values and predictions.

Usage

MAE(Y, Ypred)

Arguments

Y	a real vector with the values of the output
Ypred	a real vector with the predicted values at the same inputs

Value

a real which represents the mean of the absolute errors between the real and the predicted values:

MAE = 1/n * [|Y(x1)-Ypred(x1)|+...+|Y(xn)-Ypred(xn)|]

where xi denotes the points of the experimental design, Y the output of the computer code and Ypred the fitted model.

Author(s)

D. Dupuy

See Also

other quality criteia as RMSE and RMA.

Examples

X	<- seq(-1,1,0.1)
Y	<- 3*X + rnorm(length(X),0,0.5)
Ypred	<- 3*X
MAE(Y,Ypred)

DiceEval documentation built on July 4, 2022, 9:05 a.m.