[Corpora-List] The Implement of LDA classifier

[CARLOS EDUARDO DANTAS DE MENEZES]

Hi Wang,

I've found an e-mail to corpora list that seems to answer your question.
Regards,

Carlos Menezes

Marco Baroni to corpora
show details Mar 7
Hi everybody.

About a week ago, I asked the list for advice re tools for implementing
Topic Models.

Here is a summary of the responses I've got, for future reference -- thanks
to all those who replied!

***

Diarmuid Ó Séaghdha and Vu Hoang Cong Duy recommended the Mallet's topic
modeling library (Diarmuid also provided evidence that it scales up very
well from his experiments, and this is also what I'm going to try first):

http://mallet.cs.umass.edu/topics.php

***

Amaç Herdagdelen mentioned that the Apache Lucene Mahout project contains
functionalities for Latent Dirichlet Allocation:

http://lucene.apache.org/mahout/

***

Barbara Plank recommended Johnathan Chang's R LDA implementation (no
guarantee on scalability, but should be nice at least for prototyping):

http://cran.r-project.org/web/packages/lda/

***

Gianluca Lebani mentioned David Blei's LDA implementation and links to other
toolkits:

http://www.cs.princeton.edu/~blei/lda-c/

***

Thu Le Dieu recommended GibbsLDA++ and its Java variant, and also provided
evidence on their scalability (links to more toolkits on the first of these
pages):

http://gibbslda.sourceforge.net/
http://sourceforge.net/projects/jgibblda/

***

Thanks again to those who replied, and I hope this short survey will be also
useful to somebody else.

Regards,

Marco


-- 
Marco Baroni
Center for Mind/Brain Sciences (CIMeC)
University of Trento
http://clic.cimec.unitn.it/marco

_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora






On Tue, Jul 27, 2010 at 10:40 PM, Xiao-lin Wang
<arthur_general at sjtu.edu.cn>wrote:

>
> Hi,
>
> My research topic at present is meta-learning. I have been trying SVM as
> the meta-level classifier, and there are several popular SVM implements on
> the web. But SVM turns out to work poor mainly due to the computational
> complexity, which costs more than several weeks to learn a large meta-level
> training set.   Recently I want to try LDA classifier as the meta-level
> classifier, but I can not find a implement of LDA classifiers after spending
> hours on the web. Can any body recommend some implements of LDA?
>
>
> Many thanks.
>
>
>
> X.L. Wang
>
> arthur_general at sjtu.edu.cn
>
>
>
>
>
> _______________________________________________
> Corpora mailing list
> Corpora at uib.no
> http://mailman.uib.no/listinfo/corpora
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20100729/e6231a63/attachment.htm>
-------------- next part --------------
_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora