mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Tanton Gibbs <tanton.gi...@gmail.com>
Subject Re: Relevance score - Classification
Date Wed, 23 Nov 2011 17:23:17 GMT
Hi Faizan,

You can definitely do this, basically divide your training data into
two sets (relevant and not relevant) and then build up a classifier to
tell the difference.  However, I strongly recommend you look at a
learning to rank solution instead.  Even if you don't want to ever
show an irrelevant result, a ranking based solution that is calibrated
to a 0-1 scale is often better than a classification approach.

Can you tell us a bit more about your application and I might have
specific recommendations.

On Wed, Nov 23, 2011 at 3:21 AM, Faizan(Aroha)
<faizan.sheikh@arohalabs.net> wrote:
> Hello,
>
>
>
> We are working on using Classification as a Search.
>
>
>
> I want to compute the relevance score of the output which is generated by
> the Naive Bayes Classifier or some other classifier.
>
>
>
> Please give any guideline/hint!
>
>
>
> Thanks.
>
>
>
> --
>
> Best Regards
>
> Faizan Shaikh
> Aroha Labs (Private) Limited
>
>
>
>
>
>

Mime
View raw message