Thanks,
Clive
This might be the thread you're remembering (entitled "Use of Vector
Store RAM")
http://groups.google.com/group/semanticvectors/browse_thread/thread/9b71cf817d1dbac4
We haven't made coordinated progress on this, but Glen has some very
clear proposals, I don't know of any implementations yet.
Best wishes,
Dominic
Were you actually referring to: 1) having the documents stored in a DB
and then building SV off that; or 2) having the SV stored in the DB?
If its the former you can look at Lucene's options to index a database
- http://wiki.apache.org/lucene-java/LuceneFAQ#How_can_I_use_Lucene_to_index_a_database.3F.
Once you have the Lucene index built off the DB you can use SV as
usual.
Good lick,
Bevan
> --
> You received this message because you are subscribed to the Google Groups "Semantic Vectors" group.
> To post to this group, send email to semanti...@googlegroups.com.
> To unsubscribe from this group, send email to semanticvecto...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/semanticvectors?hl=en.
>
--
| | Bevan Koopman, PhD Candidate
|>.|<. koopman.id.au
Queensland University of Technology
Australian e-Health Research Centre