Re: [Gretl-users] Forecasting Future Trends with Gretl
On Sun, 31 Oct 2010, Anton Bletsov wrote:
I am building an ARIMA model with Gretl, and I want to forecast the
changes
in the value of my dependent variable for the period 2010-2012, based on my
previous observations. Although I added two extra observations (of course,
there are no actual values for those), I keep getting this message when I
try to make a forecast for the period:
There are no observations available for forecasting
out of sample. If you wish, you can add observations
(Data menu, Edit data), or you can shorten the sample
range over which the model is estimated (Sample menu).
Can you send me a copy of the dataset you're using, and the model
specification? From the information given, I can't say what's
happening here.
As a comparison, you might try this:
* Open the Ramanathan dataset data9-7.
* Under the Data menu, choose Add observations and select
2 additional observations
* Under the Model menu, select an ARMA(1,1) model for
the variable QNC
* Under the Analysis menu in the model window, select
Forecasts: you will find that you're offered a
dynamic forecast for the two extra observations.
This is the expected behavior.
Allin Cottrell