Highlights:
- New plots (a lot!): imputation, trend, change points, residuals, qq-plot, feature relevance, stl.
- New regressors logic in TSDatasets, Transforms and Models
- Added jupyter notebook with regressors example
- Prediction intervals visualization in plot_forecast
- Detrending could be polynomial
- Added installation instruction for M1
- Fixed TSDataset when plot method does not plot all required segments
- VotingEnsemble allows to set weights of estimator as weights of pipelines
Full changelog:
Added
- Regressors logic to TSDatasets init (#357)
- FutureMixin into some transforms (#361)
- Regressors updating in TSDataset transform loops (#374)
- Regressors handling in TSDataset make_future and train_test_split (#447)
- Prediction intervals visualization in plot_forecast (#538)
- Add plot_imputation (#598)
- Add plot_time_series_with_change_points function (#534)
- Add plot_trend (#565)
- Add find_change_points function (#521)
- Add option day_number_in_year to DateFlagsTransform (#552)
- Add plot_residuals (#539)
- Add get_residuals (#597)
- Create PerSegmentBaseModel, PerSegmentPredictionIntervalModel (#537)
- Create MultiSegmentModel (#551)
- Add qq_plot (#604)
- Add regressors example notebook (#577)
- Create EnsembleMixin (#574)
- Add option season_number to DateFlagsTransform (#567)
- Create BasePipeline, add prediction intervals to all the pipelines, move parameter n_fold to forecast (#578)
- Add stl_plot (#575)
- Add plot_features_relevance (#579)
- Add community section to README.md (#580)
- Create AbstaractPipeline (#573)
- Option "auto" to weights parameter of VotingEnsemble, enables to use feature importance as weights of base estimators (https://github.com/tinkoff-ai/etna/pull/587[](https://github.com/tinkoff-ai/etna/releases/edit/1.7.0#changed-1))
Changed
- Change the way ProphetModel works with regressors (#383)
- Change the way SARIMAXModel works with regressors (#380)
- Change the way Sklearn models works with regressors (#440)
- Change the way FeatureSelectionTransform works with regressors, rename variables replacing the "regressor" to "feature" (#522)
- Add table option to ConsoleLogger (#544)
- Installation instruction (#526)
- Update plot_forecast for multi-forecast mode (#584)
- Trainer kwargs for deep models (#540)
- Update CONTRIBUTING.md (#536)
- Rename _CatBoostModel, _HoltWintersModel, _SklearnModel (#543)
- Add logging to TSDataset.make_future, log repr of transform instead of class name (#555)
- Rename _SARIMAXModel and _ProphetModel, make SARIMAXModel and ProphetModel inherit from PerSegmentPredictionIntervalModel (#549)
- Update get_started section in README (#569)
- Make detrending polynomial (#566)
- Update documentation about transforms that generate regressors, update examples with them (#572)
- Fix that segment is string (#602)
- Make LabelEncoderTransform and OneHotEncoderTransform multi-segment (https://github.com/tinkoff-ai/etna/pull/554[](https://github.com/tinkoff-ai/etna/releases/edit/1.7.0#fixed-1))
Fixed
- Fix TSDataset._update_regressors logic removing the regressors (#489)
- Fix TSDataset.info, TSDataset.describe methods (#519)
- Fix regressors handling for OneHotEncoderTransform and HolidayTransform (#518)
- Fix wandb summary issue with custom plots (#535)
- Small notebook fixes (#595)
- Fix import Literal in plotters (#558)
- Fix plot method bug when plot method does not plot all required segments (#596)
- Fix dependencies for ARM (#599)
- [BUG] nn models make forecast without inverse_transform (#541)