Matrix/Linear Algebra Library?

[Benji Smith]

Fawzi Mohamed wrote:
> Probably if you want least square approximation (that is what you meant 
> with LSA?) some kind of direct minimizer is the way to go (but be 
> careful the problem might ill conditioned).
> Fawzi

Sorry. I meant "Latent Semantic Analysis". I'm building a table of 
synonyms from a large document repository.

There are about 100,000 documents in the repository, and the total 
corpus has about 100,000 unique words (after stemming). The LSA 
algorithm uses a matrix, where each row represents a single term, and 
each column represents a single document. Each cell contains the TF-IDF 
(Term Frequency / Inverse Document Frequency).

Performing an SVD on the matrix yields a collelation matrix, 
representing the synonymy of all terms, in all documents.

--benji