[R] use of sequence on ridge regression

[Prof Brian Ripley]

On Wed, 7 May 2008, Rodrigo Briceño wrote:

> Sorry to bother with this topic, but I'm still not clear about the
> meaning of the value set that is used on lambda values. Is there a
> correct way of doing that? My doubt is how to choose those 3 values
> that appear in the example.

I think you have not yet read the help for seq(): try printing out
seq(0,0.1,0.001).  There are not '3 values' here, but 3 arguments to a 
function.

'lambda' values start at 0 (no shrinkage), and how large you want them to 
be is problem-specific.  How many you ask for (101 here) is also 
problem-specific, but around 100 values is often used.


> Thanks.
>
> On Wed, May 7, 2008 at 9:29 AM, Prof Brian Ripley <ripley at stats.ox.ac.uk> wrote:
>> On Wed, 7 May 2008, Rodrigo Briceño wrote:
>>
>>
>>> Nice to meet you Brian. The question is about the numbers that appears
>>> after lambda (0,0, 0.1, 0.0001). I know that seq is a set of values used
>>> for testing which value fits best. But I'm not sure if I need to put
>>> whatever I think or what. I tried also with a set of 5 values and I get
>>> an error. Is there a maximum allowed?
>>>
>>
>>  So you mean the set of values of lambda?  It is just a set to be use for a
>> plot and for searching in the select() methods.
>>
>>  There is an R function called sequence(), and it is not the same as seq().
>>
>>
>>
>>
>>>
>>> plot(lm.ridge(hipcenter ~ .,seatpos, lambda = seq(0,0.1,0.001)))
>>> select(lm.ridge(hipcenter ~ ., seatpos,lambda = seq(0,0.1,0.001)))
>>>
>>> __________________________________________________________________
>>>
>>> Rodrigo Briceño
>>> Project Manager
>>> Sanigest Internacional
>>>
>>> +506  22-91-12-00 ext. 113   Oficina Costa Rica
>>> +506  22-32-08-30                  Fax
>>> +506  88-86-11-77                  Celular
>>> rbriceno at sanigest.com
>>> www.sanigest.com
>>>
>>> MSN: jbric98 at hotmail.com
>>> SKYPE: rbriceno1087
>>>
>>> _____________________
>>>
>>> This communication contains legal information which is privileged and
>>> confidential. It is for the exclusive use of the address and
>>> distribution, dissemination, copying or use by others is strictly
>>> prohibited. If you have received this communication by error, please
>>> delete the original message and e-mail us.
>>>
>>>
>>> Esta comunicación contiene información legal privilegiada y confidencial
>>> para el uso exclusivo del destinatario. La distribución, diseminación,
>>> copia u otro uso por terceras personas es estrictamente prohibida. Si
>>> usted ha recibido esta comunicación por error, le rogamos borrar el
>>> mensaje original y comunicárnoslo a esta misma dirección.
>>>
>>>
>>>
>>> Prof Brian Ripley wrote:
>>>      What do you mean by 'the sequence option'?
>>>
>>>      The authot of lm.ridge
>>>
>>>      On Wed, 7 May 2008, Rodrigo Briceño wrote:
>>>
>>>            Dear R users. I have a doubt about the use of the
>>>            sequence option on
>>>            Ridge regression. I'm trying to understand the
>>>            use of this option when
>>>            variables are highly linear correlated. I'm
>>>            running a model where the
>>>            variables HtShoes and Ht have high VIF values. My
>>>            program is written
>>>            below, but I'm not sure about the correct way of
>>>            using the sequence
>>>            option:
>>>
>>>            library (faraway)
>>>            data (seatpos)
>>>            attach (seatpos)
>>>            spos.mod <- lm(hipcenter ~ .,seatpos) summary
>>>            (spos.mod)
>>>            library(MASS)
>>>            lm.ridge(hipcenter ~ .,seatpos)
>>>            plot(lm.ridge(hipcenter ~ .,seatpos, lambda =
>>>            seq(0,0.1,0.001)))
>>>            select(lm.ridge(hipcenter ~ ., seatpos,lambda =
>>>            seq(0,0.1,0.001)))
>>>
>>>            Any advice will be appreaciated. Rodrigo B.
>>>
>>>            ______________________________________________
>>>            R-help at r-project.org mailing list
>>>            https://stat.ethz.ch/mailman/listinfo/r-help
>>>            PLEASE do read the posting guide
>>>            http://www.R-project.org/posting-guide.html
>>>            and provide commented, minimal, self-contained,
>>>            reproducible code.
>>>
>>>
>>>
>>>
>>>
>>
>>
>>  --
>>  Brian D. Ripley,                  ripley at stats.ox.ac.uk
>>  Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
>>  University of Oxford,             Tel:  +44 1865 272861 (self)
>>  1 South Parks Road,                     +44 1865 272866 (PA)
>>  Oxford OX1 3TG, UK                Fax:  +44 1865 272595
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595