Home

Yohou is a time series forecasting framework built on top of Scikit-Learn's ecosystem. It provides a unified interface for building, extending, and comparing any forecasting model, from sklearn-native reductions to statistical models or deep learning integrations and hyperparameter optimization workflows. All models share a consistent API with native DataFrame support, Scikit-Learn-based compositions, and first-class cross-validation.

• Tutorials

  ---

  Install Yohou, load a dataset, fit a forecaster, and generate your first predictions.

  Tutorials

• How-to Guides

  ---

  Step-by-step guides for panel data, exogenous features, ensembles, custom estimators, and more.

  How-to Guides

• Explanation

  ---

  Understand the fit/observe/predict lifecycle, data formats, preprocessing, and core design decisions.

  Explanation

• API Reference

  ---

  Complete documentation for every class and function across all submodules.

  API Reference

• Extensions

  ---

  Official and community extensions for hyperparameter optimization, deep learning integrations, and more.

  Extensions

• Examples

  ---

  Interactive Marimo notebooks demonstrating forecasting, metrics, visualization, and more.

  Examples

Key Features

• Polars-native: All data flows use polars.DataFrame with a mandatory "time" column.
• Sklearn-compatible: Standard fit/predict API with a consistent interface across all forecaster types.
• Reduction forecasting: Wrap any sklearn regressor or classifier and Yohou handles windowing, tabularization, and recursive prediction automatically.
• Point, interval, and class-probability forecasting: From naive baselines to conformal prediction intervals and calibrated class-probability distributions.
• Panel data: First-class support for multiple related time series via the __ column naming convention, with per-group models via LocalPanelForecaster.
• Incremental observation: Call observe() to feed new data, rewind() to roll back state, and observe_predict() to fast-forward and forecast in one step without retraining.
• Stateful transformers: All transformers are stateful and fitted separately from the forecaster, supporting incremental observation and rewind in full preprocessing pipelines.
• Composable pipelines: Decomposition pipelines, feature pipelines, feature unions, and column transformers that compose like sklearn.
• Cross-validation and model selection: Temporal splitters (ExpandingWindowSplitter, SlidingWindowSplitter) and GridSearchCV/RandomizedSearchCV with no data leakage.
• Metrics: Point, interval, and class-probability scorers with stepwise, vintagewise, componentwise, and groupwise aggregation modes.

What's New

See the Changelog for the latest release notes and updates.

License

This project is licensed under the terms of the Apache-2.0 License.

Acknowledgements

We would like to thank Evolta Technologies for their support to the project.

This project is maintained by stateful-y, an ML consultancy specializing in time series data science & engineering. If you're interested in collaborating or learning more about our services, please visit our website.