Evaluation

3 views
Skip to first unread message

Christian

unread,
Nov 6, 2010, 10:06:18 PM11/6/10
to MetaOptimize Challenge [discuss]
You write 'Edit: For any sub mission, I will post for a random subset
of vocab words each entry’s 10 related terms. I’ll then ask people to
vote blind. This is a reasonable technique for quantitative
evaluation.".
This presumes that the people who vote are representative of the
semantics encoded in your corpus. If your corpus is of general topics
then a reasonable audience probably will still achieve a good result
though.

Joseph Turian

unread,
Nov 7, 2010, 6:06:36 PM11/7/10
to metaoptimize-ch...@googlegroups.com
> This presumes that the people who vote are representative of the
> semantics encoded in your corpus. If your corpus is of general topics
> then a reasonable audience probably will still achieve a good result
> though.

True, but this is a general English usage corpus.

Reply all
Reply to author
Forward
0 new messages