Bug in MAPE?
Posted: Fri May 07, 2010 2:17 am
by Dataminer
I believe there to be a bug in the calculation of MAPE in the forecast evaluation option to .fit and .forecast. Further details are indicated in my posting in the Program repository section.
Gerald
Re: Bug in MAPE?
Posted: Fri May 07, 2010 8:47 am
by EViews Gareth
Posted a full response in your other thread, but, in short, I don't believe it is a bug, you're simply calculating MAPE incorrectly, by dividing by the forecast values, rather than dividing by the actual values.
Re: Bug in MAPE?
Posted: Mon May 10, 2010 1:53 am
by Dataminer
Error acknowledged - thanks Gareth - update in cited section.
Gerald