v0.1.0
v0.1.0 (2024-05-03)
Build
- build: remove river from default dependencies (
66a3a21)
Ci
-
ci: publish to pypi (
ec1256c) -
ci: add gh-pages (
6efd58c) -
ci: doc error (
7c1507f) -
ci: doc error (
678e0a2) -
ci: doc error (
3dda36c) -
ci: Fix doc error (
25b34df) -
ci: rename SGBT to StreamingGradientBoostedTrees (
82604a5) -
ci: change python version of github actions to 3.9 (
1a2ca36) -
ci: only upload docs on push (
f1f2416) -
ci: add macos back to
all_targets.yml(abe46f2) -
ci: fix an off by one error in version numbers (
9a8dbf1)
Documentation
-
docs: update README (
01c4a9f) -
docs: update landing pages (
ee5eee4) -
docs: fix spelling and style mistakes (
95c72fb) -
docs(SKClassifier): add docs, doctest, typehints, and minor refactor (
d85b708)
Feature
-
feat: add missing datasets and document datasets (
3ac973f) -
feat: add online smooth boost classifiers (
478bd93) -
feat: add nochange and majority class classifiers (
0c822c1) -
feat: add OzaBoost (
a08fd1b) -
feat: improve
capymoaenvironment configuration (ee96275) -
feat: add SGBT (
80f7007) -
feat(EFDT): leaf_prediction as str
Users can still use leaf_prediction as an integer (0, 1 or 2), but it can also be used as a string:
"MajorityClass": 0, "NaiveBayes": 1, "NaiveBayesAdaptive": 2 (6454179)
-
feat(regressor): add SGDRegressor using sklearn (
2e55155) -
feat(regressor): add PassiveAggressiveRegressor (
c54fe50) -
feat(base): add SKRegressor (
309aa48) -
feat(SGDClassifier): add SGDClassifier (
ec00ffd)
Fix
-
fix: update soknl and test (
336766f) -
fix: fix python 3.9 syntax error and float comparison in test (
c7b7c1b) -
fix: several updates
Updated EFDT and HoeffdingTree to
use _leaf_prediction(...) from _utils.py
Also changed dataset._util.py to
dataset.utils.py
Finally, updated the tests, there were
some issues (like EFDT_gini was using
InformationGain). (1dc6234)
Refactor
- refactor(PassiveAggressiveClassifier): use SKClassifier base class (
cb3ff18)