(MAE) A measure of forecast accuracy calculated by summing the absolute values of the individual forecast errors of the time series and dividing by the number of observations.