Some bad editing in that long post!  One statement needs a real
clarification.

In the Classification discussion section, I said:

mBria wrote
> 
>  To a lesser degree, I'd like to know I could if I wanted get the system
> to tell a a small set of clusters the new doc may fit well with with a
> "score".
> 

What I meant was:
If possible, where relevant, I'd actually even like the ability for the
system to return to me a small set (1-3) of "/most likely/ matching
classes/clusters", accompanied with a "similarity score" for each possible
match.

As I re-write this statement, I realize that I should add that I'm am also
planning to incorporate Lucene into the overall mix as well.  I mention this
because I realize that what I'm describing above is technically a
description of how Lucene works.  In the end, I think what I'm really
looking for is a way to use Mahout to get more fancy/precise about what
Lucene does with it's 'MoreLikeThis'.

Again, thanks for your help!
Mike


--
View this message in context: 
http://lucene.472066.n3.nabble.com/Some-guidance-for-this-noob-Metadata-Matching-Engine-tp3978388p3979731.html
Sent from the Mahout User List mailing list archive at Nabble.com.

Reply via email to