Re: [AutoWeka] How-To configure a specific classifier in AutoWeka

Peter Reutemann <[email protected]> Fri, 17 May 2024 09:18:33 +1200
Newsgroups gmane.comp.ai.weka
Message-ID <[email protected]>
--===============2052563247780529200==
Content-Type: multipart/alternative;
 boundary="------------y4yEVdcuB7he1ZlJfewiETkT"
Content-Language: en-US

--------------y4yEVdcuB7he1ZlJfewiETkT
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit


> I did look into the multisearch package.
>
> I plan to explore the hyperparameter tuning with a variety of 
> different classifiers.
>
> Is there an example or template that documents what property and 
> parameter range I can explore?
>

The Weka wiki has some examples:
https://waikato.github.io/weka-wiki/optimizing_parameters/#multisearch

And if you are looking for code examples, have a look at the companion 
repo of the multi-search package:
https://github.com/fracpete/multisearch-weka-package-examples

The "property" names are just the names of the properties that you see 
in the GenericObjectEditor. They can describe a path through the 
objects, if you are looking at optimizing parameters that are nested a 
bit deeper.

In terms of range for each parameter, that can only be gleaned from the 
documentation of the classifiers (and common sense), unfortunately.

Cheers, Peter
-- 
Peter Reutemann
Dept. of Computer Science
University of Waikato, Hamilton, NZ
Mobile +64 22 190 2375
https://profiles.waikato.ac.nz/peter.reutemann
http://www.data-mining.co.nz/
--------------y4yEVdcuB7he1ZlJfewiETkT
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: 7bit

<!DOCTYPE html><html><head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body>
    <br>
    <blockquote type="cite" cite="mid:CABaQXBu2iu7C8L17Lvjm15ee5U9qZMi9w6YKL4VxAoThR24QkQ@mail.gmail.com">
      <div dir="ltr">
        <div>I did look into the multisearch package.</div>
        <div><br>
        </div>
        <div>I plan to explore the hyperparameter tuning with a variety
          of different classifiers.</div>
        <div><br>
        </div>
        <div>Is there an example or template that documents what
          property and parameter range I can explore?</div>
        <br>
      </div>
    </blockquote>
    <br>
    The Weka wiki has some examples:<br>
<a class="moz-txt-link-freetext" href="https://waikato.github.io/weka-wiki/optimizing_parameters/#multisearch">https://waikato.github.io/weka-wiki/optimizing_parameters/#multisearch</a><br>
    <br>
    And if you are looking for code examples, have a look at the
    companion repo of the multi-search package:<br>
    <a class="moz-txt-link-freetext" href="https://github.com/fracpete/multisearch-weka-package-examples">https://github.com/fracpete/multisearch-weka-package-examples</a><br>
    <br>
    The &quot;property&quot; names are just the names of the properties that you
    see in the GenericObjectEditor. They can describe a path through the
    objects, if you are looking at optimizing parameters that are nested
    a bit deeper.<br>
    <br>
    In terms of range for each parameter, that can only be gleaned from
    the documentation of the classifiers (and common sense),
    unfortunately.<br>
    <br>
    Cheers, Peter<br>
    -- <br>
    <div class="moz-signature"><font size="2">Peter Reutemann<br>
        Dept. of Computer Science<br>
        University of Waikato, Hamilton, NZ<br>
        Mobile +64 22 190 2375<br>
        <a href="https://profiles.waikato.ac.nz/peter.reutemann" class="moz-txt-link-freetext">https://profiles.waikato.ac.nz/peter.reutemann</a><br>
        <a href="http://www.data-mining.co.nz/" class="moz-txt-link-freetext">http://www.data-mining.co.nz/</a></font><font size="small"><font size="2"></font></font></div>
  </body>
</html>

--------------y4yEVdcuB7he1ZlJfewiETkT--

--===============2052563247780529200==
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

_______________________________________________
Wekalist mailing list -- [email protected]
Send posts to [email protected]
To unsubscribe send an email to [email protected]
To subscribe, unsubscribe, etc., visit https://list.waikato.ac.nz/postorius/lists/wekalist.list.waikato.ac.nz
List etiquette: http://www.cs.waikato.ac.nz/~ml/weka/mailinglist_etiquette.html

--===============2052563247780529200==--