how to change Candidate Label Threshold

classic Classic list List threaded Threaded
7 messages Options
Reply | Threaded
Open this post in threaded view
|

how to change Candidate Label Threshold

曾小伟
hi ,all

In <osinski-2003-lingo.pdf>,  I find there are some LINGO's parameters,such as Candidate Label Threshold.
these parameters can influence the cluster result,

who can help me to find  where are these parameters, and how to change them,
these parameters are:
Candidate Label  Threshold
Snippet Assignment Threshold
Term Frequency Threshold
Label Similarity Threshold

Best regards
Zeng Xiaowei
Oct-20-2010







------------------------------------------------------------------------------
Download new Adobe(R) Flash(R) Builder(TM) 4
The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
Flex(R) Builder(TM)) enable the development of rich applications that run
across multiple browsers and platforms. Download your free trials today!
http://p.sf.net/sfu/adobe-dev2dev
_______________________________________________
Carrot2-developers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/carrot2-developers
Reply | Threaded
Open this post in threaded view
|

Re: [!! SPAM] ***SPAM*** how to change Candidate Label Threshold

Dawid Weiss-2
Look at the examples, they show how to set up parameters. This is
explained here:

http://download.carrot2.org/head/manual/index.html#section.integration

As for Lingo parameters, their description is here:

http://download.carrot2.org/head/manual/index.html#section.component.lingo

A good place to start tuning is by downloading the Workbench
application, where you can play with the parameters and see how they
affect the clusters in real time.

Dawid

> who can help me to find  where are these parameters, and how to change them,
> these parameters are:
> Candidate Label  Threshold
> Snippet Assignment Threshold
> Term Frequency Threshold
> Label Similarity Threshold

------------------------------------------------------------------------------
Download new Adobe(R) Flash(R) Builder(TM) 4
The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
Flex(R) Builder(TM)) enable the development of rich applications that run
across multiple browsers and platforms. Download your free trials today!
http://p.sf.net/sfu/adobe-dev2dev
_______________________________________________
Carrot2-developers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/carrot2-developers
Reply | Threaded
Open this post in threaded view
|

Re: how to change Candidate Label Threshold

Stanislaw Osinski
Administrator
In reply to this post by 曾小伟
Hello,

In addition to the previous response, the implementation of Lingo has evolved a bit from what was described in the original paper. In particular:

* Candidate Label  Threshold: does not exist any more due to the change from SVD to multiple matrix factorizations. Related parameter: http://download.carrot2.org/head/manual/#section.attribute.LingoClusteringAlgorithm.desiredClusterCountBase
* Snippet Assignment Threshold: does not exist any more due to a different document assignment strategy. Related parameter: http://download.carrot2.org/head/manual/#section.attribute.LingoClusteringAlgorithm.labelAssigner
* Term Frequency Threshold: replaced with document frequency threshold: http://download.carrot2.org/head/manual/#section.attribute.lingo.CaseNormalizer.dfThreshold
* Label Similarity Threshold: does not exist any more due to a different cluster merging algorithm: Related parameter: http://download.carrot2.org/head/manual/#section.attribute.LingoClusteringAlgorithm.clusterMergingThreshold

For a complete list of parameters of Lingo, please see:


Thanks,

Staszek

2010/10/20 曾小伟 <[hidden email]>
hi ,all

In <osinski-2003-lingo.pdf>,  I find there are some LINGO's parameters,such as Candidate Label Threshold.
these parameters can influence the cluster result,

who can help me to find  where are these parameters, and how to change them,
these parameters are:
Candidate Label  Threshold
Snippet Assignment Threshold
Term Frequency Threshold
Label Similarity Threshold

Best regards
Zeng Xiaowei
Oct-20-2010







------------------------------------------------------------------------------
Download new Adobe(R) Flash(R) Builder(TM) 4
The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
Flex(R) Builder(TM)) enable the development of rich applications that run
across multiple browsers and platforms. Download your free trials today!
http://p.sf.net/sfu/adobe-dev2dev
_______________________________________________
Carrot2-developers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/carrot2-developers



------------------------------------------------------------------------------
Download new Adobe(R) Flash(R) Builder(TM) 4
The new Adobe(R) Flex(R) 4 and Flash(R) Builder(TM) 4 (formerly
Flex(R) Builder(TM)) enable the development of rich applications that run
across multiple browsers and platforms. Download your free trials today!
http://p.sf.net/sfu/adobe-dev2dev
_______________________________________________
Carrot2-developers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/carrot2-developers
Reply | Threaded
Open this post in threaded view
|

Re: [!! SPAM] ***SPAM*** how to change Candidate Label Threshold

waleedAzmy
In reply to this post by Dawid Weiss-2
But what if I want to set these parameters in the XML file called "algorithm-lingo-attributes.xml"

where can I found this file?

note: I'm using the source code of Carrot2 not just the examples and Eclipse IDE,
Reply | Threaded
Open this post in threaded view
|

Re: [!! SPAM] ***SPAM*** how to change Candidate Label Threshold

Dawid Weiss-2
ctrl-shift-r in Eclipse will give you the default locations. It is in
component-suites project.

Dawid

On Thu, Oct 21, 2010 at 12:54 PM, waleedAzmy <[hidden email]> wrote:

>
> But what if I want to set these parameters in the XML file called
> "algorithm-lingo-attributes.xml"
>
> where can I found this file?
>
> note: I'm using the source code of Carrot2 not just the examples and Eclipse
> IDE,
> --
> View this message in context: http://carrot2-users-and-developers-forum.607571.n2.nabble.com/how-to-change-Candidate-Label-Threshold-tp5653531p5658189.html
> Sent from the Carrot2 Users and Developers Forum mailing list archive at Nabble.com.
>
> ------------------------------------------------------------------------------
> Nokia and AT&T present the 2010 Calling All Innovators-North America contest
> Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
> http://p.sf.net/sfu/nokia-dev2dev
> _______________________________________________
> Carrot2-developers mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/carrot2-developers
>

------------------------------------------------------------------------------
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev
_______________________________________________
Carrot2-developers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/carrot2-developers
Reply | Threaded
Open this post in threaded view
|

Re: [!! SPAM] ***SPAM*** how to change Candidate Label Threshold

waleedAzmy
I found this file in two locations.

\carrot2\core\carrot2-component-suites\suites\suites
and
\carrot2\core\carrot2-component-suites\tmp\eclipse\suites

which one of them? I try to change this attribute...

      <attribute key="MultilingualClustering.defaultLanguage">
        <value type="org.carrot2.core.LanguageCode" value="ARABIC"/>
      </attribute>

But It still cluster Arabic documents as English.... I think it does not reflect any changes in the code...
Reply | Threaded
Open this post in threaded view
|

Re: [!! SPAM] ***SPAM*** how to change Candidate Label Threshold

Dawid Weiss-2
> \carrot2\core\carrot2-component-suites\suites\suites

This one, the other one comes from the output compilation path.

> which one of them? I try to change this attribute...
>
>      <attribute key="MultilingualClustering.defaultLanguage">
>        <value type="org.carrot2.core.LanguageCode" value="ARABIC"/>
>      </attribute>
>
> But It still cluster Arabic documents as English.... I think it does not
> reflect any changes in the code...

This is the default language taken in the absence of language marker
in the input. Which tool are you using again -- the workbench, the
DCS, are you clustering from the code directly? Document sources
should set the language flag properly, for example Microsoft's Bing
document source (BingDocumentSource) has a property called
"BingDocumentSource.market" which will query Bing with the appropriate
culture information and set the language for clustering automatically.

If you cluster from code, look at this complete example:
ClusteringNonEnglishContent
https://fisheye3.atlassian.com/browse/carrot2/branches/stable/applications/carrot2-examples/src/org/carrot2/examples/clustering/ClusteringNonEnglishContent.java?r=trunk

Dawid

------------------------------------------------------------------------------
Nokia and AT&T present the 2010 Calling All Innovators-North America contest
Create new apps & games for the Nokia N8 for consumers in  U.S. and Canada
$10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing
Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store
http://p.sf.net/sfu/nokia-dev2dev
_______________________________________________
Carrot2-developers mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/carrot2-developers