SPSSX Discussion

Question about ridge regression

Classic

List

Threaded

9 messages Options

David Greenberg

Question about ridge regression

About a dozen years ago or more I used the syntax below to enter a correlation matrix into SPSS, and then estimate a ridge regression on it. Now I am using version 21, and it doesn't work. Could someone telll me how to modify the syntax so that it will run? Thank you, David Greenberg, Sociology Department, New York University
     matrix data variables=ftp unemp lic hgr clear hom
     /contents = n_scalar corr/format=upper.
begin data
    13
      1   .293 .569 .702 -.974 .964
             1    -.167 -.033 -.306 .210
                      1      .904 -.555 .726
                                 1   -.683   .816
                                          1    -.968
                                                   1
end data.
begin matrix 'include 'P:\spsswin\ridgereg.sps'.
ridgereg dep=hom/enter=ftp unemp lic hgr clear
   /start=0/stop=1/inc=.05.
end matrix.

David Marso

Re: Question about ridge regression

Administrator

David,
You are proceeding under the false assumption that RidgeReg supports MATRIX input.
Note towards the beginning of the macro

correlations variables=!dep !enter /missing=listwise/matrix out(*).
set errors on results listing .

YOU could possibly modify the macro to bypass the correlations command and read the current active file.

What you have here would never have worked.

Where does this begin matrix .. end matrix come from?
You can't just make up commands and expect them to work (there is no BEGIN MATRIX in SPSS).

David

David Greenberg wrote

About a dozen years ago or more I used the syntax below to enter a
correlation matrix into SPSS, and then estimate a ridge regression on it.
Now I am using version 21, and it doesn't work. Could someone telll me how
to modify the syntax so that it will run? Thank you, David Greenberg,
Sociology Department, New York University
matrix data variables=ftp unemp lic hgr clear hom
/contents = n_scalar corr/format=upper.
begin data
13
1 .293 .569 .702 -.974 .964
1 -.167 -.033 -.306 .210
1 .904 -.555 .726
1 -.683 .816
1 -.968
1
end data.
begin matrix 'include 'P:\spsswin\ridgereg.sps'.
ridgereg dep=hom/enter=ftp unemp lic hgr clear
/start=0/stop=1/inc=.05.
end matrix.

Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me.
---
"Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis."
Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?"

Jon K Peck

Re: Question about ridge regression

In reply to this post by David Greenberg

Note that ridge regression is available as a built-in procedure in the Categories option (along with lasso and elastic net).

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: 720-342-5621

From: David Greenberg <[hidden email]>
To: [hidden email],
Date: 10/27/2013 09:32 PM
Subject: [SPSSX-L] Question about ridge regression
Sent by: "SPSSX(r) Discussion" <[hidden email]>

David Marso

Re: Question about ridge regression

Administrator

And for those who lack the $ for categories?
I suspect it would take you (Jon) about 5-10 minutes to add support for MATRIX input, multiple data sets (rather than saving the data to a temp file), and make the output prettier.
AFAICT: I am the very last person to have touched that code probably back in 1996 (I was still in teksport) ?
---

Jon K Peck wrote

Note that ridge regression is available as a built-in procedure in the
Categories option (along with lasso and elastic net).

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: 720-342-5621

From: David Greenberg <[hidden email]>
To: [hidden email],
Date: 10/27/2013 09:32 PM
Subject: [SPSSX-L] Question about ridge regression
Sent by: "SPSSX(r) Discussion" <[hidden email]>

About a dozen years ago or more I used the syntax below to enter a
correlation matrix into SPSS, and then estimate a ridge regression on it.
Now I am using version 21, and it doesn't work. Could someone telll me how
to modify the syntax so that it will run? Thank you, David Greenberg,
Sociology Department, New York University
matrix data variables=ftp unemp lic hgr clear hom
/contents = n_scalar corr/format=upper.
begin data
13
1 .293 .569 .702 -.974 .964
1 -.167 -.033 -.306 .210
1 .904 -.555 .726
1 -.683 .816
1 -.968
1
end data.
begin matrix 'include 'P:\spsswin\ridgereg.sps'.
ridgereg dep=hom/enter=ftp unemp lic hgr clear
/start=0/stop=1/inc=.05.
end matrix.

Yang, Hongwei

Re: Question about ridge regression

In reply to this post by Jon K Peck

Hi Jon,

It is good to know that SPSS provides the ridge regression option along with LASSO and ELASTIC NET. I did not know this before reading this message. So, it is the menu under Analyze/Regression/Optimal Scaling (CATREG)?

Thanks a lot!!

Hongwei Yang, University of Kentucky

From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jon K Peck
Sent: Monday, October 28, 2013 8:18 AM
To: [hidden email]
Subject: Re: Question about ridge regression

Jon K Peck

Re: Question about ridge regression

Yes, see the Regularization subdialog.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
phone: 720-342-5621

From: "Yang, Hongwei" <[hidden email]>
To: Jon K Peck/Chicago/IBM@IBMUS, "[hidden email]" <[hidden email]>,
Date: 10/28/2013 10:51 AM
Subject: RE: Question about ridge regression

Hi Jon,

It is good to know that SPSS provides the ridge regression option along with LASSO and ELASTIC NET. I did not know this before reading this message. So, it is the menu under Analyze/Regression/Optimal Scaling (CATREG)?

Thanks a lot!!

Hongwei Yang, University of Kentucky

From: SPSSX(r) Discussion [[hidden email]] On Behalf Of Jon K Peck
Sent: Monday, October 28, 2013 8:18 AM
To: [hidden email]
Subject: Re: Question about ridge regression

Note that ridge regression is available as a built-in procedure in the Categories option (along with lasso and elastic net).

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@...
phone: 720-342-5621

From: David Greenberg <dg4@...>
To: [hidden email],
Date: 10/27/2013 09:32 PM
Subject: [SPSSX-L] Question about ridge regression
Sent by: "SPSSX(r) Discussion" <[hidden email]>

About a dozen years ago or more I used the syntax below to enter a correlation matrix into SPSS, and then estimate a ridge regression on it. Now I am using version 21, and it doesn't work. Could someone telll me how to modify the syntax so that it will run? Thank you, David Greenberg, Sociology Department, New York University
matrix data variables=ftp unemp lic hgr clear hom
/contents = n_scalar corr/format=upper.
begin data
13
1 .293 .569 .702 -.974 .964
1 -.167 -.033 -.306 .210
1 .904 -.555 .726
1 -.683 .816
1 -.968
1
end data.
begin matrix 'include 'P:\spsswin\ridgereg.sps'.
ridgereg dep=hom/enter=ftp unemp lic hgr clear
/start=0/stop=1/inc=.05.
end matrix.

parvin

Re: Question about ridge regression

In reply to this post by Jon K Peck

CONTENTS DELETED

The author has deleted this message.

Jon Peck

Re: Question about ridge regression

RR is available via the OPTREG command. There is also a RR macro around somewhere.

On Wed, Jan 22, 2020 at 7:10 PM parvin <[hidden email]> wrote:

Hello, What is the syntax of ridge regression in spss?

--
Sent from: http://spssx-discussion.1045642.n5.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

Jon K Peck
[hidden email]

===================== 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

parvin

Re: Question about ridge regression

CONTENTS DELETED

The author has deleted this message.