Re: Different R^2 values with same variables
- From: Jerry Dallal <gdallal@xxxxxxxxxxxxxxxxxxxxxxxx>
- Date: Tue, 28 Mar 2006 06:44:32 -0500
pekkajarvela@xxxxxxxxx wrote:
If I run linear regression with R^2 selection (in SAS) and get some
sets of variables with R^2 values. Then I took the best set of
variables (with largest R^2) and run regression with these variables
without any selection. I get different (worse) R^2 value with these
same variables than in regression with R^2 selection! How can that be
so?
Pekka
I would guess that missing values are involved. It's likely that some cases were missing values for variables defining the global search and were "set aside". However. they were not missing any of the variables in your final model and were now included.
.
- Follow-Ups:
- Re: Different R^2 values with same variables
- From: pekkajarvela
- Re: Different R^2 values with same variables
- References:
- Different R^2 values with same variables
- From: pekkajarvela
- Different R^2 values with same variables
- Prev by Date: Variance as central moment
- Next by Date: Re: multicollinearity in regression
- Previous by thread: Different R^2 values with same variables
- Next by thread: Re: Different R^2 values with same variables
- Index(es):
Relevant Pages
|
|
