Mlforecast Versions Save

Bug Fixes

• fix RollingStd typo @jmoralez (#276)

Enhancement

• Improve error message "Found missing inputs in X_df" @MarcoGorelli (#273)
• use backtest_splits from utilsforecast @jmoralez (#271)

New Features

• support lag transformations from coreforecast @jmoralez (#265)
• add feature_engineering module @jmoralez (#261)
• add as_numpy argument @jmoralez (#249)
• add polars support @jmoralez (#241)

Breaking Change

• remove deprecated arguments @jmoralez (#240)
• remove dynamic_dfs argument @jmoralez (#239)

Bug Fixes

• deterministic column order @jmoralez (#262)
• fix inverse transforms for fitted values when series were dropped @jmoralez (#255)
• Fix distributed cv @jmoralez (#254)
• add packaging to dependencies @jmoralez (#235)

Documentation

• add as_numpy guide @jmoralez (#258)
• add analyzing models and custom training how-to guides @jmoralez (#236)

Enhancement

• keep df order in cv @jmoralez (#257)
• handle short series exception @jmoralez (#256)
• support polars dataframe in TimeSeries.update @jmoralez (#252)
• issue warning instead of error for short series in cv @jmoralez (#247)
• ensure lags are positive integers @jmoralez (#232)

Full Changelog: https://github.com/Nixtla/mlforecast/compare/v0.10.0...v0.11.0

Breaking Change

• remove differences argument @jmoralez (#215)

Bug Fixes

• fix X_df slices @jmoralez (#228)

Documentation

• move distributed API reference to quickstart @jmoralez (#229)
• extract how-to guides from API reference @jmoralez (#224)
• Feature/electricity load forecasting (PJM) tutorial using MLForecast @uumami (#208)
• Prediction intervals for machine learning models @Naren8520 (#196)

Bug Fixes

• support fitted inverse transform in LocalStandardScaler @jmoralez (#206)
• fix fitted_values methods @jmoralez (#198)

Enhancement

• raise error when max_horizon and models_ don't match @jmoralez (#204)

New Features

• add forecast_fitted_values method @jmoralez (#190)
• support integer refit in cross_validation @jmoralez (#189)
• add GlobalSklearnTransformer @jmoralez (#187)

New Features

• support predicting a subset of series @jmoralez (#183)

Enhancement

• raise informative error when interpreting dynamic features as static @jmoralez (#182)

Enhancement

• faster MLForecast.preprocess @jmoralez (#179)
• deprecate dynamic_dfs argument in favor of X_df @jmoralez (#176)
• improve static_features_ definition @jmoralez (#175)

Bug Fixes

• fix TimeSeries.update method @jmoralez (#173)
• fix static_features order @jmoralez (#174)

New Features

• Add LocalStandardScaler @jmoralez (#171)

Enhancement

• make argument names compatible with other nixtla libraries @jmoralez (#166)

Bug Fixes

• fix keep_last_n with target_transforms @jmoralez (#171)

New Features

• add cross validation fitted values @jmoralez (#164)
• allow id_col in static_features @jmoralez (#161)

Enhancement

• raise error for wrong frequency in cross_validation @jmoralez (#160)
• Add new release drafter @FedericoGarza (#146)
• Add new issue template @FedericoGarza (#145)