Significance testing using approximate randomisation and bootstrap resampling

wikilinks / conll03_nel_eval

Python evaluation scripts for AIDA-formatted CoNLL data

Apache License 2.0

20 stars 4 forks source link

Significance testing using approximate randomisation and bootstrap resampling #27

Closed jnothman closed 10 years ago

jnothman commented 10 years ago

TODO: unit tests

jnothman commented 10 years ago

@benhachey if you can point me to data files to compare, I'll test this out.

jnothman commented 10 years ago

Although I've still not got a definitive answer on bootstrap resampling (I might got borrow Noreen's text on it later), and haven't implemented unit tests, I'll be merging this lest it get more out of date with the rest of the repo.

jnothman commented 10 years ago

This still requires nicer output.