You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@uima.apache.org by Christoph Büscher <ch...@neofonie.de> on 2008/01/17 15:43:57 UTC

Anybody with experiences using the Mayo Weka/UIMA Integration

Hi,

does anyone on this list have experience with the Mayo Weka/UIMA Integration 
(http://informatics.mayo.edu/text/index.php?page=weka) or other Libraries that 
enable UIMA components to use machine learning algorithms?

I would be especially interested how easy/difficult the integration of Weka 
classes and algorithms was, how active the project seems to be and if anybody 
did anything else than simple test applications with it.

I'm also especially interested in UIMA components that annotate/transform text 
documents with feature representations (like TF/IDF, bag of words) usually used 
in classification in machine learning, e.g. to later train an SVM with it. Any 
ideas about existing code out there would also help a great deal.

Thanks,

Christoph

RE: Anybody with experiences using the Mayo Weka/UIMA Integration

Posted by Olivier Terrier <ol...@temis.com>.
You should give a try to the Jena University JNET annotator here
http://www.julielab.de/content/view/117/174/
It uses machine learning-based named entities extraction
 
Best 
 
Olivier

________________________________

From: Christoph Büscher [mailto:christoph.buescher@neofonie.de]
Sent: Thu 1/17/2008 3:43 PM
To: uima-user@incubator.apache.org
Subject: Anybody with experiences using the Mayo Weka/UIMA Integration



Hi,

does anyone on this list have experience with the Mayo Weka/UIMA Integration
(http://informatics.mayo.edu/text/index.php?page=weka) or other Libraries that
enable UIMA components to use machine learning algorithms?

I would be especially interested how easy/difficult the integration of Weka
classes and algorithms was, how active the project seems to be and if anybody
did anything else than simple test applications with it.

I'm also especially interested in UIMA components that annotate/transform text
documents with feature representations (like TF/IDF, bag of words) usually used
in classification in machine learning, e.g. to later train an SVM with it. Any
ideas about existing code out there would also help a great deal.

Thanks,

Christoph

RE: Anybody with experiences using the Mayo Weka/UIMA Integration

Posted by Pascal Coupet <pa...@temis.com>.
The Mayo Weka integration is based on the old IBM UIMA framework according to the web site and it will not be compatible "out of the box" with the Apache version. 
The Jena JNET annotator pointed by Olivier works with the Apache version. 

Pascal  
 

-----Original Message-----
From: Christoph Büscher [mailto:christoph.buescher@neofonie.de] 
Sent: Thursday, January 17, 2008 9:44 AM
To: uima-user@incubator.apache.org
Subject: Anybody with experiences using the Mayo Weka/UIMA Integration

Hi,

does anyone on this list have experience with the Mayo Weka/UIMA Integration 
(http://informatics.mayo.edu/text/index.php?page=weka) or other Libraries that 
enable UIMA components to use machine learning algorithms?

I would be especially interested how easy/difficult the integration of Weka 
classes and algorithms was, how active the project seems to be and if anybody 
did anything else than simple test applications with it.

I'm also especially interested in UIMA components that annotate/transform text 
documents with feature representations (like TF/IDF, bag of words) usually used 
in classification in machine learning, e.g. to later train an SVM with it. Any 
ideas about existing code out there would also help a great deal.

Thanks,

Christoph