search

xiaoyangren / dkpro-core-asl

Automatically exported from code.google.com/p/dkpro-core-asl
0 stars 0 forks source link

Integrate Stanford sentiment analyzer #424

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
Stanford CoreNLP 3.3.1 and above provides a module for sentiment analysis of 
sentences. It gives five scores for very negative, negative, neutral, positive, 
and very positive (a distribution).

It can be easily integrated to stanfordnlp-gpl, but at the moment, it is quite 
slower as it runs the whole Stanford pipeline independent of previous 
annotations done in DKPro (we leave it as open issue for now).

There are two patches attached:

dkpro-asl:
* new submodule de.tudarmstadt.ukp.dkpro.core.api.sentiment-asl with typesystem 
(not a general typesystem for sentiment, only one exact matching the scores 
provided by Stanford analyzer; can be refactored if needed)

dkpro-gpl:
* de.tudarmstadt.ukp.dkpro.core.stanfordnlp.StanfordSentimentAnalyzer and a 
simple test case

Original issue reported on code.google.com by ivan.hab...@gmail.com on 18 Jul 2014 at 11:56

Attachments:

GoogleCodeExporter commented 9 years ago 
There is a method called 
de.tudarmstadt.ukp.dkpro.core.stanfordnlp.util.TreeUtils.createStanfordTree(Anno
tation, TreeFactory) that should help to transform a DKPro Core constituent 
tree into the Stanford representation. 

Maybe that helps, but maybe it takes us only half the way. Do you know what 
information the Stanford sentiment analyzer requires? Only the parse tree or 
more?

Original comment by richard.eckart on 21 Jul 2014 at 8:07

GoogleCodeExporter commented 9 years ago 
The dependency on the models artifact seems to be required right now because 
the sentiment component does not use the DKPro Core model provisioning 
mechanism.

This mechanism should be used and the dependency on the models should be 
removed:

        <dependency>
            <groupId>edu.stanford.nlp</groupId>
            <artifactId>stanford-corenlp</artifactId>
            <version>3.4</version>
            <classifier>models</classifier>
        </dependency>

Original comment by richard.eckart on 30 Jul 2014 at 8:52

GoogleCodeExporter commented 9 years ago

Original comment by richard.eckart on 14 Aug 2014 at 10:05

GoogleCodeExporter commented 9 years ago

Original comment by richard.eckart on 2 Oct 2014 at 9:32