ARIMA model

[juliby]

I am trying to forecast the price of oil using arima model. I have obtained the ARIMA (1,1,0) as the appropriate model. Please can someone tell me how to go about estimating the model parameters and forecasting using eviews.i am really confused on how to do it. Please can someone help me estimate this data. its from may 1987 to march 2009. i'm trying to forecast with eviews and having lots of sleepless night with.

monthly data 1987m05-2009m03.xls

(26 KiB) Downloaded 518 times

i really will be very grateful if someone can help solve my problem.
thanks

[startz]

Basically, you just need to include an AR(1) term on the right of a regression, using the D() function on the left. What have you tried so far?

[juliby]

I have tried it and got a forecast. I'm not sure how to obtian the confidence interval for the forecast.

[gkavcak]

Basically, you just need to include an AR(1) term on the right of a regression, using the D() function on the left. What have you tried so far?

Hi Startz,

I have a question related with your answer. I have the same problem. I have a ARIMA(1,0,0). I estimated an equation like this...

D(euro) = c + AR(1) + e

but I didnt have an forecasted data. Eviews says something like I am out of range...

I suppose there is something I missed but I could not find out what. If you can help me, it will be appreciated.

Thanks and Regards,

[startz]

Basically, you just need to include an AR(1) term on the right of a regression, using the D() function on the left. What have you tried so far?

Hi Startz,

I have a question related with your answer. I have the same problem. I have a ARIMA(1,0,0). I estimated an equation like this...

D(euro) = c + AR(1) + e

but I didnt have an forecasted data. Eviews says something like I am out of range...

I suppose there is something I missed but I could not find out what. If you can help me, it will be appreciated.

Thanks and Regards,

To get help with this kind of question, you need to provide much more information. Try posting a picture of your estimated equation, the error message you got, and your workfile.

[gkavcak]

Startz,

Actually I found the solution but I have a different problem related with an ARIMA model. Briefly, I have a time series from 2009 to 2012 with monthly freq. I firstly transformed it to logaritmics. Then I checked the correlogram (pic 1) but the graph shows me a non-stationary form. Then I created 1st difference of the values and looked at the correlogram (pic 2)which looks like a stationary form.

In the 2nd stage I generated a model. On the lef hand side I used log of my time series and on the right hand side I used the AR(1). When I checked the results of the model (pic 3) everything seems ok except "Inverted AR Roots". I says "Estimated AR process is nonstationary. I am not sure what I have to do at this step. If you can look at the printscreens for further information and share your thought, it will be appreciated.

Thanks in advance and regards.

[startz]

Use the difference of the log on your regression.

[gkavcak]

Use the difference of the log on your regression.

Thanks startz.