EViews.com

I am trying to do a forecast evaluation where the forecast is stored as a series and the observed values are also a series.

I follow the instructions listed here: http://www.eviews.com/help/helpintro.html#page/content/series-Forecast_Evaluation.html, yet I always obtain unusual results.

I am not sure what I am doing wrong. My observed series is titled RETURNSRQ, and the forecast series is titled VOLFORECASTGARCH. The sample range is 2015 to 2018 and i set the training sample to the same( I am not exactly sure about what to set the training sample too.). I filled out this dialog box like this:


This is the result I obtain:


I expected the graph to look like this:


The statistics that come with the evaluation are also nonsensical.

I am not sure what I am doing wrong. Any help would be greatly appreciated. For some reason I can't attach my workfile.

Forecast evaluation is for evaluation multiple forecasts of the same series. It looks like you are only evaluating a single series. Thus most of the statistics, and the averaging methods are nonsensical.

EViews Gareth wrote:Forecast evaluation is for evaluation multiple forecasts of the same series. It looks like you are only evaluating a single series. Thus most of the statistics, and the averaging methods are nonsensical.

So how do I obtain forecast evaluation statistics like root mean square error? How can I compare between forecast values and real observations?

You can use the forecast evaluation view for that, but don't include an averaging method.

EViews Gareth wrote:You can use the forecast evaluation view for that, but don't include an averaging method.

I followed your instructions and obtained this:


I don't understand why its not plotting the full evaluation sample.

Hard to say without having the data.

EViews Gareth wrote:Hard to say without having the data.

Sorry, I thought I had attached it. I have attached it now. Cheers.

Works as expected for me. Ensure your copy of EViews is up to date.

EViews Gareth wrote:2018-07-13_090917.png

I just want to say thanks for all your help. Updating the software did the trick. Sorry for taking up your time.
Cheers.