Templating sentence construction and inserting data is completed, although not yet for all the positions, but only for what's more popular (QBs, RBs, WRs, TEs, and scoring).
Simple adjective-replacement in sentence diversity is working. Attaching clauses to main sentence body is a little impractical based solely on probability, so I'm thinking tuples based on key words in the main body would make more sense, i.e.:
Elaboration on scoring.
Providing context for high-stat plays.
Still trying to find a way to isolate events (other than TDs and major yards) without putting too many additions into the content, such as:
Failed 4th down conversions at crucial moments.
Bigger defensive plays.
Still playing around with pandas. Hopefully something will hit me soon.
Other to do's:
Use NLTK's entity extraction to ensure subject relevance in generated content.
Find a way to properly evaluate the validity of NLTK's entity extraction.
Templating sentence construction and inserting data is completed, although not yet for all the positions, but only for what's more popular (QBs, RBs, WRs, TEs, and scoring).
Simple adjective-replacement in sentence diversity is working. Attaching clauses to main sentence body is a little impractical based solely on probability, so I'm thinking tuples based on key words in the main body would make more sense, i.e.:
Still trying to find a way to isolate events (other than TDs and major yards) without putting too many additions into the content, such as:
Other to do's: