top | item 45512260

(no title)

svg7 | 4 months ago

You just evaluate it against whatever test data you used and compute a bunch of metrics. You decide to use the model, if "bad things" happen at an acceptable enough rate.

discuss

order

No comments yet.