in reply to Natural Language Sentence Production
So, there's a lot of literature on natural language... Assuming that you want a bit quicker ramp-up than a few months of reading, I'd do a search on citeseer for data on this, since someone has probably written a paper.
Failing that, you could scan corpora for sentences that contain the words on the magnets, or you could generate a model from corpora (such as a bayesian one), then constrain your model to only sentences containing those words, and choose sentences with the highest probability.
Failing that, you could scan corpora for sentences that contain the words on the magnets, or you could generate a model from corpora (such as a bayesian one), then constrain your model to only sentences containing those words, and choose sentences with the highest probability.
|
---|
Replies are listed 'Best First'. | |
---|---|
Re^2: Natural Language Sentence Production
by Camel_thirst (Friar) on Sep 23, 2005 at 10:56 UTC |
In Section
Seekers of Perl Wisdom