v1.7.0
New Features
- support integer refit in cross_validation @jmoralez (#731)
- support forecast_fitted_values in distributed @jmoralez (#732)
- use environment variable to get id as column in outputs @jmoralez (#721)
- support different column names for ids, times and targets @jmoralez (#718)
- add mstl_decomposition function @jmoralez (#701)
- [FEAT] Ability to save and load StatsForecast @akmalsoliev (#667)
- support caching all jitted functions @jmoralez (#651)
Bug Fixes
- fix incorrect computation of n in ARIMA @StatKumar (#708)
- fix predict_insample docstrings @jmoralez (#713)
- fix arima var_coef @jmoralez (#706)
- feat: _calculate_sigma() explicitly handles series with 1 data point (n=0) @cparmet (#699)
- [Feat] Robustified crossvalidation @kdgutier (#668)
- support exogenous features in distributed @jmoralez (#638)
- fix MSTL where trend forecaster supports level @jmoralez (#625)
Documentation
- run how-to docs @jmoralez (#730)
- run getting-started docs @jmoralez (#726)
- add exogenous features support to models table @jmoralez (#714)
- use numpy style in all docstrings @jmoralez (#695)
- Fix Minimal example and ARIMA family heading @yibenhuang (#694)
- fix equations in documentation @jmoralez (#697)
- Isolated import cells [Issue #653] @patrick-dd (#691)
- MLFlow Tutorial @kvnkho (#676)
- Add Mintlify automations @hahnbeelee (#671)
- Correct broken urls in docs @DaveParr (#660)
- Fixed error in GARCH tutorial that corresponds to issue #629 @MMenchero (#650)
- Restoring example using Conformal Prediction on StatsForecast @kvnkho (#643)
- arima model user guide @Naren8520 (#570)
Dependencies
- remove pyarrow dependency for polars @jmoralez (#705)
- [UPDATE] Updated Docker & requirements.txt @akmalsoliev (#672)
- [CHORE] Updating
requirements.txt
@akmalsoliev (#666) - test support python 3.11 @msfidelis (#425)