search

edkinsgael / airhead-research

Automatically exported from code.google.com/p/airhead-research
0 stars 0 forks source link

DocumentSpace reformulation #36

Closed GoogleCodeExporter closed 8 years ago

GoogleCodeExporter commented 8 years ago 
DocumentSpace should not exist

Instead there should be a DocumentVectorBuilder, possibly located under
common, which takes a SemanticSpace and a few options.  Then, given a
document, it will request term vectors, which are assumed to be weighted
correctly, and sums/averages/other magic the term vectors to project the
document in the semantic space.

For an option regarding vector weighting, it should have at least term
frequency and no weighting.

This will change how esa works.  Esa will simply be a SSpace, and to
process documents, it can be passed into a DocumentVectorBuilder.

Original issue reported on code.google.com by FozzietheBeat@gmail.com on 6 Nov 2009 at 12:26

GoogleCodeExporter commented 8 years ago

Original comment by FozzietheBeat@gmail.com on 6 Nov 2009 at 5:38