Add voting learners by desilinguist · Pull Request #665 · EducationalTestingService/skll

added 30 commits

June 9, 2020 14:21
- This is useful since it can the be used by both regular learners as well as the voting learners.
- Add docstring for `cross_validate()`
- Add `learning_curve()`.
- Use refactored code where possible.
# Conflicts:
#	skll/learner/__init__.py
#	skll/learner/utils.py
# Conflicts:
#	skll/learner/utils.py
# Conflicts:
#	tests/test_output.py

@desilinguist

- Use position arguments instead of `.args` and `.kwargs` accessors which do not seem to be supported on Python 3.7.

@desilinguist

@desilinguist

- Set BINDIR since we are now running tests without activating the conda environment.
- Set nose options as environment variables to make command shorter
- Use `travis_wait` for the longest running test job to avoid early termination.

@desilinguist

@desilinguist

@desilinguist

@desilinguist

@desilinguist

@desilinguist

@desilinguist

@desilinguist

- Update run_experiment documentation to include detailed description of `VotingClassifier` and `VotingRegressor`. Also add entry for `save_votes` configuration field.
- Add new votingconfiguration files for voting learners for Iris and Boston examples.
- Update contributing page for readability and fix links.
- Remove all top-level imports from documentation pages and tutorial notebook.
- Update Learner API documentation to include `VotingLearner` class and improved sub-headings.
- Update docstrings for voting learners to make them more readable.
- Update sphinx configuration (year and imports
-

@desilinguist

@desilinguist

mulhod

@desilinguist

- Update API and experiment tests to handle this new functionality.

mulhod

@desilinguist

aoifecahill