Page 1 of 1
How to estimate the model with sub samples?
Posted: Sat May 10, 2014 12:44 am
by researcher84
Code: Select all
OPEN DATA "C:\data.xls"
CALENDAR(M) 1986:01
DATA(FORMAT=XLS,NOLABELS,ORG=COLUMNS) 1990:01 2013:07 cpi ip inflation
*
Is what I have above right? I assume this will run my model on the subsample 1990:01 to 2013:07. If not, how to run different subsamples with same program and data file ? my data has no lables, hopefully what I have above right. Thanks alot!
Re: How to estimate the model with sub samples?
Posted: Sat May 10, 2014 7:57 am
by TomDoan
Since you didn't attach the data set, I have no idea if it's correct.
If the file has dates on it (you can have dates and not labels, or vice versa), then you would probably want just
OPEN DATA "C:\data.xls"
CALENDAR(M) 1990:01
DATA(FORMAT=XLS,NOLABELS,ORG=COLUMNS) 1990:01 2013:07 cpi ip inflation
so if you don't want the 1986-1989 data, you just leave it out. If the file has no dates, then you have to do
OPEN DATA "C:\data.xls"
CALENDAR(M) 1986:01
DATA(FORMAT=XLS,NOLABELS,ORG=COLUMNS) 1986:01 2013:07 cpi ip inflation
If there are no dates on the data file, DATA assumes the first line corresponds to entry 1 on your CALENDAR. If your data set has no dates or labels, I would strongly suggest that you read it in once, and write a new version that includes both.
If you're stuck with the version with no dates and labels (if, for instance, it's being updated by someone else), then read it with the second set of instructions and just use the start and end parameters on a LINREG or whatever to restrict the range.