SPSSX Discussion

pls regression

Classic

List

Threaded

7 messages Options

vlad simion

pls regression

Hi all,

could anyone help me with some clearings on the PLS Regression, the
algorithm, what are the assumptions for running it and so on. And where can
I find some materials on pls regression?

Thank you in advance.

Vlad.

--
Vlad Simion
Data Analyst
Tel: +40 720130611

Dan Zetu

Re: pls regression

First, I am not aware of any SPSS capability to do PLS regression. SAS has
proc pls that does PLS regression, not to be confused with the PLS approach
for latent variable modeling. If you have SAS, then its documentation has a
rather good description of the procedure and what it does.

Although I personally haven't seen one, any chemometrics book should contain
at least some brief description of the PLS regression procedure, as it is
this field in which PLS has found many applications. Also, SAS researchers
have published papers in the past on PLS regression (they should be publicly
available on SAS website).

Hope this helps,

Dan

>From: vlad simion <[hidden email]>
>Reply-To: vlad simion <[hidden email]>
>To: [hidden email]
>Subject: pls regression
>Date: Thu, 29 Jun 2006 09:15:49 +0300
>
>Hi all,
>
>could anyone help me with some clearings on the PLS Regression, the
>algorithm, what are the assumptions for running it and so on. And where can
>I find some materials on pls regression?
>
>Thank you in advance.
>
>Vlad.
>
>--
>Vlad Simion
>Data Analyst
>Tel: +40 720130611

Anthony Babinec

Re: pls regression

In reply to this post by vlad simion

If you have access to backissues of Technometrics,
there's a nice article by Frank and Friedman on
PLS and related approaches:

Frank, Ildiko and Friedman, Jerome. 1993. "A Statistical
View of Some Chemometrics Regression Tools." Technometrics,
Vol. 35, No. 2, pp. 109-135.

This paper also has several discussants.

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
vlad simion
Sent: Thursday, June 29, 2006 1:16 AM
To: [hidden email]
Subject: pls regression

Hi all,

could anyone help me with some clearings on the PLS Regression, the
algorithm, what are the assumptions for running it and so on. And where can
I find some materials on pls regression?

Thank you in advance.

Vlad.

--
Vlad Simion
Data Analyst
Tel: +40 720130611

vlad simion

Re: pls regression

In reply to this post by Dan Zetu

Thank you all,

I think I am a little bit confused and I am not sure I really know what is
appropriate for what I need.
I have some data with a lot of missings where the independent variables are
highly correlated and as far as I understand the PLS regression (partial
least squares - if I get it right) would be more fitted than multivariate
regression or GLM, because of the multicolinearity.
I hope it cleares a little what my problem is, because for me....

Thank you again.

Vlad.

On 6/29/06, Dan Zetu <[hidden email]> wrote:

>
> First, I am not aware of any SPSS capability to do PLS regression. SAS has
> proc pls that does PLS regression, not to be confused with the PLS
> approach
> for latent variable modeling. If you have SAS, then its documentation has
> a
> rather good description of the procedure and what it does.
>
> Although I personally haven't seen one, any chemometrics book should
> contain
> at least some brief description of the PLS regression procedure, as it is
> this field in which PLS has found many applications. Also, SAS researchers
> have published papers in the past on PLS regression (they should be
> publicly
> available on SAS website).
>
> Hope this helps,
>
> Dan
>
>
> >From: vlad simion <[hidden email]>
> >Reply-To: vlad simion <[hidden email]>
> >To: [hidden email]
> >Subject: pls regression
> >Date: Thu, 29 Jun 2006 09:15:49 +0300
> >
> >Hi all,
> >
> >could anyone help me with some clearings on the PLS Regression, the
> >algorithm, what are the assumptions for running it and so on. And where
> can
> >I find some materials on pls regression?
> >
> >Thank you in advance.
> >
> >Vlad.
> >
> >--
> >Vlad Simion
> >Data Analyst
> >Tel: +40 720130611
>
>
>

--
Vlad Simion
Data Analyst
Tel: +40 720130611

Kooij, A.J. van der

Re: pls regression

use Ridge Regression, which is included in SPSS as a macro:

INCLUDE 'SPSS installdir\Ridge regression.sps'.
RIDGEREG DEP=dependent variable /ENTER = independent variables
[/START={0**}] [/STOP={1**}] [/INC={0.05**}]
{value} {value} {value }
[ /K=value]
SPSS installdir is the directory in which SPSS is installed

Regards,
Anita
________________________________

From: SPSSX(r) Discussion on behalf of vlad simion
Sent: Fri 30/06/2006 08:39
To: [hidden email]
Subject: Re: pls regression

Thank you all,

I think I am a little bit confused and I am not sure I really know what is
appropriate for what I need.
I have some data with a lot of missings where the independent variables are
highly correlated and as far as I understand the PLS regression (partial
least squares - if I get it right) would be more fitted than multivariate
regression or GLM, because of the multicolinearity.
I hope it cleares a little what my problem is, because for me....

Thank you again.

Vlad.

On 6/29/06, Dan Zetu <[hidden email]> wrote:

--
Vlad Simion
Data Analyst
Tel: +40 720130611

**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
**********************************************************************

vlad simion

Re: pls regression

In reply to this post by vlad simion

Thank you very much!

Vlad.

On 6/30/06, Dan Zetu <[hidden email]> wrote:

>
>
> Vlad,
>
> PLS is best suited for the case when you have many variables and small
> sample sizes and does not assume multivariate normality. It does also have
> an intrinsic way of dealing with multicollinearity by explaining both the
> covariance among the IVs and the association between IVs and DVs.
>
> But, if you have a lot of missing data, a Chaid-type analysis may be more
> appropriate for your case.
>
> Dan
>
> >From: "vlad simion" <[hidden email]>
> >To: "Dan Zetu" <[hidden email]>
> >CC: [hidden email]
> >Subject: Re: pls regression
> >Date: Fri, 30 Jun 2006 09:39:53 +0300
> >
> >Thank you all,
> >
> >I think I am a little bit confused and I am not sure I really know what
> is
> >appropriate for what I need.
> >I have some data with a lot of missings where the independent variables
> are
> >highly correlated and as far as I understand the PLS regression (partial
> >least squares - if I get it right) would be more fitted than multivariate
> >regression or GLM, because of the multicolinearity.
> >I hope it cleares a little what my problem is, because for me....
> >
> >Thank you again.
> >
> >Vlad.
> >
> >
> >On 6/29/06, Dan Zetu <[hidden email]> wrote:
> >>
> >>First, I am not aware of any SPSS capability to do PLS regression. SAS
> has
> >>proc pls that does PLS regression, not to be confused with the PLS
> >>approach
> >>for latent variable modeling. If you have SAS, then its documentation
> has
> >>a
> >>rather good description of the procedure and what it does.
> >>
> >>Although I personally haven't seen one, any chemometrics book should
> >>contain
> >>at least some brief description of the PLS regression procedure, as it
> is
> >>this field in which PLS has found many applications. Also, SAS
> researchers
> >>have published papers in the past on PLS regression (they should be
> >>publicly
> >>available on SAS website).
> >>
> >>Hope this helps,
> >>
> >>Dan
> >>
> >>
> >> >From: vlad simion <[hidden email]>
> >> >Reply-To: vlad simion <[hidden email]>
> >> >To: [hidden email]
> >> >Subject: pls regression
> >> >Date: Thu, 29 Jun 2006 09:15:49 +0300
> >> >
> >> >Hi all,
> >> >
> >> >could anyone help me with some clearings on the PLS Regression, the
> >> >algorithm, what are the assumptions for running it and so on. And
> where
> >>can
> >> >I find some materials on pls regression?
> >> >
> >> >Thank you in advance.
> >> >
> >> >Vlad.
> >> >
> >> >--
> >> >Vlad Simion
> >> >Data Analyst
> >> >Tel: +40 720130611
> >>
> >>
> >>
> >
> >
> >--
> >Vlad Simion
> >Data Analyst
> >Tel: +40 720130611
>
>
>

--
Vlad Simion
Data Analyst
Tel: +40 720130611

Feinstein, Zachary

Re: pls regression

In reply to this post by vlad simion

Now the big question is... Can PLS be done with SPSS as easily as it
can with SAS? If so, would anyone have any code to send along?

Zachary

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
vlad simion
Sent: Friday, June 30, 2006 8:49 AM
To: [hidden email]
Subject: Re: pls regression

Thank you very much!

Vlad.

On 6/30/06, Dan Zetu <[hidden email]> wrote:
>
>
> Vlad,
>
> PLS is best suited for the case when you have many variables and small

> sample sizes and does not assume multivariate normality. It does also
> have an intrinsic way of dealing with multicollinearity by explaining
> both the covariance among the IVs and the association between IVs and
DVs.

>
> But, if you have a lot of missing data, a Chaid-type analysis may be
> more appropriate for your case.
>
> Dan
>
> >From: "vlad simion" <[hidden email]>
> >To: "Dan Zetu" <[hidden email]>
> >CC: [hidden email]
> >Subject: Re: pls regression
> >Date: Fri, 30 Jun 2006 09:39:53 +0300
> >
> >Thank you all,
> >
> >I think I am a little bit confused and I am not sure I really know
> >what
> is
> >appropriate for what I need.
> >I have some data with a lot of missings where the independent
> >variables
> are
> >highly correlated and as far as I understand the PLS regression
> >(partial least squares - if I get it right) would be more fitted than

> >multivariate regression or GLM, because of the multicolinearity.
> >I hope it cleares a little what my problem is, because for me....
> >
> >Thank you again.
> >
> >Vlad.
> >
> >
> >On 6/29/06, Dan Zetu <[hidden email]> wrote:
> >>
> >>First, I am not aware of any SPSS capability to do PLS regression.
> >>SAS
> has
> >>proc pls that does PLS regression, not to be confused with the PLS
> >>approach for latent variable modeling. If you have SAS, then its
> >>documentation
> has
> >>a
> >>rather good description of the procedure and what it does.
> >>
> >>Although I personally haven't seen one, any chemometrics book should

> >>contain at least some brief description of the PLS regression
> >>procedure, as it
> is
> >>this field in which PLS has found many applications. Also, SAS
> researchers
> >>have published papers in the past on PLS regression (they should be
> >>publicly available on SAS website).
> >>
> >>Hope this helps,
> >>
> >>Dan
> >>
> >>
> >> >From: vlad simion <[hidden email]>
> >> >Reply-To: vlad simion <[hidden email]>
> >> >To: [hidden email]
> >> >Subject: pls regression
> >> >Date: Thu, 29 Jun 2006 09:15:49 +0300
> >> >
> >> >Hi all,
> >> >
> >> >could anyone help me with some clearings on the PLS Regression,
> >> >the algorithm, what are the assumptions for running it and so on.
> >> >And
> where
> >>can
> >> >I find some materials on pls regression?
> >> >
> >> >Thank you in advance.
> >> >
> >> >Vlad.
> >> >
> >> >--
> >> >Vlad Simion
> >> >Data Analyst
> >> >Tel: +40 720130611
> >>
> >>
> >>
> >
> >
> >--
> >Vlad Simion
> >Data Analyst
> >Tel: +40 720130611
>
>
>

--
Vlad Simion
Data Analyst
Tel: +40 720130611