Page 1 of 1
Creating new group with negative dependant variable
Posted: Mon Jun 04, 2018 5:49 pm
by selvi
Hi,
I would like to modify an existing group which contains 5 time-series, of which the first time series is the dependant variable.
I want to create a new group which only includes those rows of the old group, for which the dependant variable is less than zero.
Then save the adjusted data set as a new group (so I don't want to use sample).
Re: Creating new group with negative dependant variable
Posted: Mon Jun 04, 2018 7:25 pm
by EViews Gareth
Groups contain series, and series all have to have the same number of observations. Thus you can't create a group that contains a subset of of rows of another group. Groups all have the same number of rows.
Re: Creating new group with negative dependant variable
Posted: Wed Jun 06, 2018 3:06 am
by selvi
Thank you Gareth,
What I meant was: suppose a group has 5 series, 1st one being a dependant variable, and 30 rows.
Suppose, I only want the rows with a dependant variable that is negative. Dependant variables with negative values can be found in rows 16-30.
Then the new group contains a 15x5 matrix of the data that can be found in rows 16-30, and columns 1-5.
How do I do this?
Thanks
Re: Creating new group with negative dependant variable
Posted: Wed Jun 06, 2018 5:48 am
by startz
Re: Creating new group with negative dependant variable
Posted: Wed Jun 06, 2018 6:12 am
by selvi
Thank you Startz
However, I would like to create a new group, which is a subset of an existing group.
I would need a function so that the new group contains data from the old group, but only with those rows where the dependant variable is negative.
Amy help would be greatly appreciated.
Re: Creating new group with negative dependant variable
Posted: Wed Jun 06, 2018 6:28 am
by startz
In EViews, groups are just collections of series and all series in a workfile page are the same length. You can exclude observations from processing by using the smpl command or by setting elements equal to NA.
If you want, I suppose you could do
Code: Select all
smpl 16 30
yy = y
xx1 = x1
xx2 = x2
xx3 = x3
xx4 = x4
smpl @all
group gg yy xx1 xx2 xx3 xx4