Improve testing of supervised learning and elm.pipeline.steps

ContinuumIO / elm

Phase I & part of Phase II of NASA SBIR - Parallel Machine Learning on Satellite Data

http://ensemble-learning-models.readthedocs.io

43 stars 27 forks source link

Improve testing of supervised learning and elm.pipeline.steps #171

Closed PeterDSteinberg closed 6 years ago

PeterDSteinberg commented 7 years ago

While working on #166 I found a bug in elm.pipeline.steps.DropNaRows (PR fix #170 ).

The following need to be better tested as part of the elm/earthio refactor:

supervised learning (this bug was discovered at this time because we have lower test/docs/examples coverage of supervised learning than unsupervised)
elm.pipeline.steps.DropNaRows and other elm.pipeline.steps . Note, as described in #149, the elm.sample_util modules are moving to earthio.filters, a new subpackage.

PeterDSteinberg commented 6 years ago

Fixed by #192 . #192 is a large increase in test coverage (~1800-2000 tests are being run in there) and a reduction in the amount of code used for testing utilities.