Why cant SemanticVectors be used directly ?

[Sandeep]

hi guys,

Noob alert: I dont have a theoretical background in AI or NLP

I was looking to do something similar in the (loosely) financial
industry, where I wanted to determine strongly correlated technical
words and use them as search suggestions. Which is why, this project
is very interesting to me.

I came across the SemanticVectors package (http://code.google.com/p/
semanticvectors/) which seems to be a parallelizable technique for "
thesaurus generation, knowledge representation, and concept matching".

I'm having a bit of an understanding problem - why couldnt this be
used out of the box ?

Thanks!
-Sandeep

[Joseph Turian]

Sandeep,

I would be very interested to see a SemanticVectors based approach.
This seems like a natural solution to the problem which would work out
of the box. I haven't used this code yet though.

Joseph