Plot a graph that shows how much adding additional training data improves labeller accuracy.

nasa-petal / PeTaL-labeller

The PeTaL labeler labels journal articles with biomimicry functions.

https://petal-labeller.readthedocs.io/en/latest/

The Unlicense

6 stars 3 forks source link

Plot a graph that shows how much adding additional training data improves labeller accuracy. #40

Closed bruffridge closed 3 years ago

bruffridge commented 3 years ago

It would be very useful to show the "learning curve" for our training efforts. So start with a subset of papers, train, see results. Add more papers and repeat. Keep adding and repeating and see the plot. Is accuracy improving ? How rapidly?

bruffridge commented 3 years ago

Done. https://github.com/nasa-petal/PeTaL-labeller/tree/main/auto-labeler/MATCH#effect-of-training-dataset-size-on-match-precision