ARIMA Modeling

[sreeja.gupta]

Hi,

I am trying to build an ARIMA model in Eviews with GDP data. The data is quarterly and I have 69 observations. On trying to understand whether the data has seasonality I am having the following problem:

For the level series, the ACF is showing increasing spikes in intervals of 4, like, s, 4s, 8s...; the PACF has significantly large 1st, 3rd and 5th spikes and then dies down gradually.

For the differenced series, the ACF is showing exponential decay in seasonal lags of 2 periods; i.e. significant spikes at s, 2s, 4s...but dying down gradually. PACF has significant 2nd, 3rd and 4th spike and then dies down.

Augmented Dickey Fuller says that the differenced series is stationary.

How do I proceed with this data.

Any help will be much appreciated.

Thanks in advance.

[sreeja.gupta]

Hi,

Is there a way by which after running an ARIMA model on a transformed series (log/differenced) I can forecast for the original series.

When I do a Proc Forecast/ Make Model I get forecasts for the transformed series only.

Thanks,
Sreeja

[EViews Gareth]

How did you enter your dependent variable?