lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tommaso Teofili <tommaso.teof...@gmail.com>
Subject Re: Reg Lucene Naive Bayesian classifier.
Date Tue, 15 Jan 2013 10:46:04 GMT
2013/1/15 VIGNESH S <vigneshklncit@gmail.com>

> Hi All,
>
> Thanks for your replies..
>
> Actually I am trying to classify the email mail data in to categories
> and also spam mails .. I have tried clustering but it is not useful
> since we can not control categories.
>
> I am looking for a light weight implementation which can be used in
> mobiles in client side.
>
> I thought Lucene Naive Bayesian Would be useful...
>
> Please Suggest me Whether classifying emails will be done using this
> Lucene Naive Bayesian or any other Lucene Classifiers..
>

You could actually use one of the existing ones (naive bayes or nearest
neighbor) or even implement a new one (just implement the Classifier
interface [1]) if you already have enough labeled data in your index (one
field containing mail test and one field containing assigned category).
To use those just call Classifier#train method to train the classifier and
Classifier#assignClass to assign a class/category to a new text.
If your task is just spam detection IMHO one of the above should be enough,
if you have also to assign different categories depending on proper
semantics then I'd recommend use some other library which is more focused
for that purpose like Apache Mahout, Apache OpenNLP, etc..

My 2 cents,
Tommaso


[1] :
http://svn.apache.org/repos/asf/lucene/dev/trunk/lucene/classification/src/java/org/apache/lucene/classification/Classifier.java


>
>
> Thanks and Regards
> Vignesh Srinivasan
>
>
> On Mon, Jan 14, 2013 at 7:23 PM, VIGNESH S <vigneshklncit@gmail.com>
> wrote:
> > Hi,
> >
> > Anyone Used the Naive Bayesian Classifier?
> >
> > It will be really helpful if some one Can  post how to use the
> > classifiers in Lucene ..
> >
> > --
> > Thanks and Regards
> > Vignesh Srinivasan
> > 9739135640
>
>
>
> --
> Thanks and Regards
> Vignesh Srinivasan
> 9739135640
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-user-help@lucene.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message