how to find best regression model

[jimjohn]

if i have a multiple linear regression model i want to perform, but i want to see what filters of other variable sin my data set will provide the best model. is there a way for me to use spss to just find the best model like that, or can i only do this through trial and error and just trying different filters and then doing regression on those filters. thanks.

[ViAnn Beadle]

What do you mean by filter here? Are you varying the variables in the
equation or the cases? If the former, the notion of best subset is somewhat
controversial among statisticians and if the mean the latter, that's called
cooking the data. Best is especially tenuous if your variables are highly
correlated.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
jimjohn
Sent: Monday, March 10, 2008 8:35 AM
To: [hidden email]
Subject: how to find best regression model

if i have a multiple linear regression model i want to perform, but i want
to
see what filters of other variable sin my data set will provide the best
model. is there a way for me to use spss to just find the best model like
that, or can i only do this through trial and error and just trying
different filters and then doing regression on those filters. thanks.
--
View this message in context:
http://www.nabble.com/how-to-find-best-regression-model-tp15950784p15950784.
html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[SR Millis-3]

First, you should avoid using stepwise variable
selection.

Frank Harrell's book, "Regression Modeling
Strategies," provides a systemic method in model
building.  Unfortunately, it is difficult to implement
these more modern model building strategies in SPSS,
eg, restricted cubic splines, calibration and
validation via the bootstrap, shrinkage/penalized
methods, etc.  Fortunately, these methods can be
implemented in R, which is free.

In addition, Bayesian model averaging can be done in
R.

LASSO and LAR are now available in SAS.

Scott Millis



--- jimjohn <[hidden email]> wrote:

> if i have a multiple linear regression model i want
> to perform, but i want to
> see what filters of other variable sin my data set
> will provide the best
> model. is there a way for me to use spss to just
> find the best model like
> that, or can i only do this through trial and error
> and just trying
> different filters and then doing regression on those
> filters. thanks.

Scott R Millis, PhD, MEd, ABPP (CN,CL,RP), CStat
Professor & Director of Research
Dept of Physical Medicine & Rehabilitation
Wayne State University School of Medicine
261 Mack Blvd
Detroit, MI 48201
Email:  [hidden email]
Tel: [phone]
Fax: [phone]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD