[prev in list] [next in list] [prev in thread] [next in thread]
List: wekalist
Subject: Re: [Wekalist] Regress out nuisance attributes before attribute selection
From: Paul <paul.m.nz () gmail ! com>
Date: 2013-01-29 22:59:07
Message-ID: CAASwtFfc_6DAGZnmj_NyHJpXCtt-x4fZ7jt-giSqXvAL6+dxKw () mail ! gmail ! com
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
I don't believe that there is a filter that does linear detrends.
There are filters however that add attributes by applying a mathematical
expression from existing attributes which may be able to be used.
/paul
On Wed, Jan 30, 2013 at 9:12 AM, Stefano Diciotti <stefano.diciotti@unifi.i=
t
> wrote:
> Paul,
>
> thank you for your useful suggestion. Another question to fully solve my =
problem: is there a filter which applies a linear detrend to all attributes=
considering some attributes (nuisance variables) as independent variables =
(to remove the effect of nuisance variables before classification)?
>
> Thanks,
>
> Stefano
>
> >I think what your looking for is another filter before the attribute
> >selected classifier. Have a look at this:
>
> >weka.classifiers.meta.FilteredClassifier -F
> >"weka.filters.unsupervised.attribute.Remove -R 1,2" -W
> >weka.classifiers.meta.AttributeSelectedClassifier -- -E
> >"weka.attributeSelection.CfsSubsetEval " -S
> >"weka.attributeSelection.BestFirst -D 1 -N 5" -W weka.classifiers.trees.=
J48
> >-- -C 0.25 -M 2
>
> >Using the 'weka.filters.unsupervised.attribute.Remove' will take out
> >attributes (in this case attribute indicies 1 & 2) prior to passing the
> >data set to the attr selected classifier.
>
> >HTH,
> >/paul
>
>
> On Tue, Jan 29, 2013 at 2:37 AM, Stefano Diciotti <stefano.diciotti@unifi=
.it
>
> wrote:
>
> Dear Weka users,
>
> Is it possible, using Weka, to regress out nuisance attributes (e.g. age
> and gender) and take the residuals (and excluding the nuisance attributes=
),
> perform attribute selection and classification? This should be done in a
> cross-validation framework (the regression and attribute selection
> performed on the training set and applied to the test set).
>
> Without the regression of nuisance attributes, the
> "AttributeSelectedClassifier" executes attribute selection and
> classification in cross-validation, but I do not find a proper filter to
> regress out nuisance attributes.
>
> Thanks in advance,
>
> Stefano
>
>
> --
>
>
>
> ****
>
> _______________________________________________
> Wekalist mailing list
> Send posts to: Wekalist@list.scms.waikato.ac.nz
> List info and subscription status:
> https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist
> List etiquette:
> http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html
>
>
[Attachment #5 (text/html)]
I don't believe that there is a filter that does linear detrends.<br><br>There \
are filters however that add attributes by applying a mathematical expression from \
existing attributes which may be able to be used.<br><br> /paul<br><br><br><div \
class="gmail_quote">On Wed, Jan 30, 2013 at 9:12 AM, Stefano Diciotti <span \
dir="ltr"><<a href="mailto:stefano.diciotti@unifi.it" \
target="_blank">stefano.diciotti@unifi.it</a>></span> wrote:<br> <blockquote \
class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid \
rgb(204,204,204);padding-left:1ex">
<div text="#000000" bgcolor="#FFFFFF">
<pre>Paul,
thank you for your useful suggestion. Another question to fully solve my problem: is \
there a filter which applies a linear detrend to all attributes considering some \
attributes (nuisance variables) as independent variables (to remove the effect of \
nuisance variables before classification)?
Thanks,
Stefano
>I think what your looking for is another filter before the attribute
>selected classifier. Have a look at this:
>weka.classifiers.meta.FilteredClassifier -F
>"weka.filters.unsupervised.attribute.Remove -R 1,2" -W
>weka.classifiers.meta.AttributeSelectedClassifier -- -E
>"weka.attributeSelection.CfsSubsetEval " -S
>"weka.attributeSelection.BestFirst -D 1 -N 5" -W \
weka.classifiers.trees.J48 >-- -C 0.25 -M 2
>Using the 'weka.filters.unsupervised.attribute.Remove' will take out
>attributes (in this case attribute indicies 1 & 2) prior to passing the
>data set to the attr selected classifier.
>HTH,
>/paul
On Tue, Jan 29, 2013 at 2:37 AM, Stefano Diciotti <<a \
href="mailto:stefano.diciotti@unifi.it" target="_blank">stefano.diciotti@unifi.it</a> \
</pre><div><div class="h5"> <blockquote type="cite" style>
<pre>wrote:
</pre>
</blockquote>
<blockquote type="cite" style>
<pre> Dear Weka users,
Is it possible, using Weka, to regress out nuisance attributes (e.g. age
and gender) and take the residuals (and excluding the nuisance attributes),
perform attribute selection and classification? This should be done in a
cross-validation framework (the regression and attribute selection
performed on the training set and applied to the test set).
Without the regression of nuisance attributes, the
"AttributeSelectedClassifier" executes attribute selection and
classification in cross-validation, but I do not find a proper filter to
regress out nuisance attributes.
Thanks in advance,
Stefano
--
</pre>
</blockquote>
<div><br>
<div align="left"><font>
<p class="MsoNormal" style="margin:0in 0in 0pt" align="left"><span \
style="font-size:7.5pt;color:gray;font-family:Arial"><font><br> \
<u></u><u></u></font></span></p> </font></div>
</div>
</div></div></div>
<br>_______________________________________________<br>
Wekalist mailing list<br>
Send posts to: <a href="mailto:Wekalist@list.scms.waikato.ac.nz">Wekalist@list.scms.waikato.ac.nz</a><br>
List info and subscription status: <a \
href="https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist" \
target="_blank">https://list.scms.waikato.ac.nz/mailman/listinfo/wekalist</a><br> \
List etiquette: <a href="http://www.cs.waikato.ac.nz/%7Eml/weka/mailinglist_etiquette.html" \
target="_blank">http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html</a><br>
<br></blockquote></div><br>
_______________________________________________
Wekalist mailing list
Send posts to: Wekalist@list.scms.waikato.ac.nz
List info and subscription status: https://list.scms.waikato.ac.nz/mailman/=
listinfo/wekalist
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.=
html
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic