Subscribe Bookmark RSS Feed

Standard error on predicted values


Community Member


Jan 9, 2013


I have used an ARMA(2, 2) model to forecast 8 steps ahead and was curious in how JMP10 calculates the standard error for these predictive values. I understand that the standard error for the in-sample values is sqrt(sum of squared erros/df) = 0.008246259, but how do JMP calculate the out-of-sample / forecasted values (equal to 0.0083182608; 0.0087230318; 0.0091141695; 0.0092120035 - see screenshot)?

Thank in advance, and I really appreciate your help!