[
https://issues.apache.org/jira/browse/TIKA-2016?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15970216#comment-15970216
]
ASF GitHub Bot commented on TIKA-2016:
--------------------------------------
thammegowda commented on issue #169: TIKA-2016 Sentiment Analysis Parser
Contributed by amensiko and thammegowda
URL: https://github.com/apache/tika/pull/169#issuecomment-294326796
1. The categorical datasets links in Datasets page is broken, we have to fix
it too. http://irds.usc.edu/SentimentAnalysisParser/datasets.html
2. The Models page has no models linked, the paths mentioned in the web page
doesnt exists in the repo.
http://irds.usc.edu/SentimentAnalysisParser/models.html. I am hoping the
Categorical analysis (or multi class classifier in simple words) should work
simply by changing the `modelPath` parameter in tika config XML (Please give me
a link to model, I will test it :+1: ). By default, it uses netflix review
dataset binary label model, but changing it is as easy as editing the tika
config XML.
3. The OpenNLP API has changed. I upgraded from 1.5.x to 1.6.0, it went
fine. However, the latest is 1.7.2, in this version, they renamed `SentimentME`
as `DocumentCagetegorizerME`, thus the existing _sentiment analysis_ models
won't work and we are stuck up here until we regenerate models.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
> A parser that combines Apache OpenNLP and Apache Tika and provides facilities
> for automatically deriving sentiment from text.
> -----------------------------------------------------------------------------------------------------------------------------
>
> Key: TIKA-2016
> URL: https://issues.apache.org/jira/browse/TIKA-2016
> Project: Tika
> Issue Type: New Feature
> Components: parser
> Reporter: Anastasija Mensikova
> Assignee: Chris A. Mattmann
> Labels: analysis, gsoc2016, memex, parser, sentiment
> Fix For: 1.15
>
>
> A new project that implements a parser that uses Apache OpenNLP and Apache
> Tika to perform Sentiment Analysis.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
