turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Forecasting
- /
- PROC ARIMA

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

02-04-2016 01:19 PM

Hi Everyone,

I'm doing some ARIMA forecasting and am finding the need to take the first difference of my variable of interest so I can get a stationary series. My variable of interest is hospital admissions. I have two questions:

1) If I have a numeric covariate, that is also time related, such as membership over time, do I need to take the first difference of that covariate?

2) If the answer to #1 is yes, how do I accomplish that in a PROC ARIMA statement?

Thank you!

**proc** **arima** data=arima_input3 plots=all;

identify var=admit_count (**1**)

crosscorr=(memberships)

nlag=**53**;

estimate q=**2** input=(memberships);

forecast lead=**106** interval=week id=admit_date out=results;

**run**;

Accepted Solutions

Solution

02-11-2016
11:25 PM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

02-05-2016 10:11 AM

It is usually a good practice to take the same diff of the crosscorr variables as the dependent var but it's up to the modeler to decide if diffing the crosscorr var is needed. You can accomplished it by adding the desirable diffs right after the crosscorr var (see below)

**proc** **arima** data=arima_input3 plots=all;

identify var=admit_count (**1**)

crosscorr=(memberships(1))

nlag=**53**;

estimate q=**2** input=(memberships);

forecast lead=**106** interval=week id=admit_date out=results;

**run**;

All Replies

Solution

02-11-2016
11:25 PM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

02-05-2016 10:11 AM

It is usually a good practice to take the same diff of the crosscorr variables as the dependent var but it's up to the modeler to decide if diffing the crosscorr var is needed. You can accomplished it by adding the desirable diffs right after the crosscorr var (see below)

**proc** **arima** data=arima_input3 plots=all;

identify var=admit_count (**1**)

crosscorr=(memberships(1))

nlag=**53**;

estimate q=**2** input=(memberships);

forecast lead=**106** interval=week id=admit_date out=results;

**run**;

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

02-11-2016 11:25 PM

Thank you!