v0.1.6

What's Changed

• add "method" option to averaging_opt by @nyanp in #90
• docs(readme): Fix Typo sklean -> sklearn by @ryanrussell in #91
• docs(readme): Grammar fixes by @ryanrussell in #92
• Rename 'sklearn' to 'scikit-learn' by @momijiame in #96
• drop supporting python 3.6/3.7 by @nyanp in #99
• Fix ci errors by @nyanp in #103
• support user-defined function(ex: numpy.sum) by @wakame1367 in #105
• Add python 3.9, 3.10 and 3.11 by @nyanp in #107
• v0.1.6 by @nyanp in #108

New Contributors

• @ryanrussell made their first contribution in #91
• @momijiame made their first contribution in #96

Full Changelog: v0.1.5...v0.1.6

v0.1.5

This release contains changes to support scikit-learn 1.0. #89

v0.1.4

This release contains the following updates & fix. Thanks @Wakamezake !

New module

• Add nyaggle.feature.aggregation function #81 by @Wakamezake

Bug Fix

• Remove deprecated arguments in optuna integration #80
• Support newer versions of transformers #77 #82 by @Wakamezake
• Support lightgbm 3.0.0 #83 #84

Tests

• Add test for find_best_lgbm_parameter #76 by @Wakamezake
• Add weekly test #79 by @Wakamezake

v0.1.3

This release contains the following bugfix & updates. Thanks @daikikatsuragawa and @tenajima !

Bug Fix

• ignore MLflowException in log_param, log_metric #73
• Fix error in run_experiment with callable fit_params #74

Docs / Others

• Update README.md according to markdownlint #68 by @daikikatsuragawa
• fix docstring of split.py #70 by @tenajima

v0.1.2

This release contains works from 2 contributors. Thanks @Wakamezake and @harupy !

Overview:

New module

• Add nyaggle.ensemble #60

Enhancement

nyaggle.validation

• Refactor StratifiedGrouKFold (remove dependency to ubelt, add type check etc) #40 #41 #43 #45 by @Wakamezake

nyaggle.experiment

• Replace overwrite option with if_exists #47
• Add inherit_experiment parameter #50
• Add log_metrics for Experiment #59 by @harupy
• Add log_dict for Experiment #65

Tests & dev

• Improve .gitignore #51 #64 @harupy
• Add PR trigger #52
• Refactor test (add options, use fixture) #53 #54 #54 @harupy

v0.1.1

• add Python3.8 support #32
• add Skip, Nth validation #33
• remove predict_proba parameter from cross_validate #37
• remove dependency to gbdt packages #38
• add StratifiedGroupKFold validation #39
• add tutorial documents
• change default value of overwrite parameter in save_feature from True to False

v0.1.0

• replace experiment_gbdt with run_experiment #27
• support xgboost #21
• add cached_feature decorator #28
• fix categorical feature handling #22

v0.0.11

• Fix module nyaggle.feature module couldn't be found from pip package

v0.0.10

• with_id option to support atmaCup#3 format :)

v0.0.9

v0.0.9