We're also thinking about implementing something similar to LSI within
ActiveMath which is lucene-powered where both formulae and text
searching would benefit of the latent-semantic-similarity. I've been
refrained of doing "exactly this" at least since LSI is patented. This
might also be a reason why there's no implementation in Lucene's sandbox.
Have you looked at other vector-based approaches which are not exactly LSI ?
Have you looked at InfoMap NLP ?
paul
[EMAIL PROTECTED] wrote:
Yeah, its Latent Semantic Indexing
On 12/1/05, gekkokid <[EMAIL PROTECTED]> wrote:
sorry have to ask - whats LSI - " latent semantic indexing"?
_gk
----- Original Message -----
From: "Lorenzo Viscanti" <[EMAIL PROTECTED]>
To: <java-user@lucene.apache.org>; <[EMAIL PROTECTED]>
Sent: Thursday, December 01, 2005 12:02 AM
Subject: Re: Lucene + LSI
It depends on the kind of implementation you are thinking of.
You can use Lucene to create the inputs to the LSI, and then use them in
your own system. I've written that code and it works, for searches and
clustering.
But if you are figuring out an LSI enhanced Lucene search system (based on
a
specific Similarity implementation?), I have to tell you that It wouldn't
be
so easy to integrate LSI into the Lucene's APIs.
Lorenzo
On 12/1/05, Chandana <[EMAIL PROTECTED]> wrote:
Have any one implemented LSI in Lucene?
Kindly let me know how hard/easy it is.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
