.. -- mode: rst --

|gitter|_ |Binder|_

.. |gitter| image:: https://img.shields.io/gitter/room/alan-turing-institute/sktime?logo=gitter .. _gitter: https://gitter.im/sktime/community

.. |binder| image:: https://mybinder.org/badge_logo.svg .. _Binder: https://mybinder.org/v2/gh/sktime/sktime-tutorial-pydata-amsterdam-2020/main?filepath=notebooks

This tutorial was written for sktime version 0.4.2. For more up-to-date notebooks visit sktime's online documentation <https://www.sktime.org/en/latest/how_to_get_started.html>__.

Introduction to Machine Learning with Time Series

This is the repository for the "Introduction to Machine Learning with Time Series" code breakfast at PyData Festival Amsterdam 2020.

You can watch the video here: https://www.youtube.com/watch?v=Wf2naBHRo8Q

You'll learn about:

• Machine learning with time series,
• How to tell apart different learning problems (or tasks) that arise in a temporal data setting,
• How to build machine learning models to solve these tasks (using sktime <https://github.com/alan-turing-institute/sktime>_ and scikit-learn <https://scikit-learn.org/stable/>_),
• How to contribute to sktime.

We assume familiarity with the standard tabular machine learning setting covered by scikit-learn <https://scikit-learn.org/stable/>_, but no prior experience of working with time series.

How to get started

You can either

• run the notebooks on Binder_ without having to install anything,
• pip install sktime <https://alan-turing-institute.github.io/sktime/installation.html>_ and clone <https://help.github.com/en/github/creating-cloning-and-archiving-repositories/cloning-a-repository>_ this repository to run the notebooks locally. This requires a working Python installation (e.g. Anaconda distribution <https://docs.anaconda.com/anaconda/install/>) with Jupyter notebooks <https://jupyter.org/install>.

How to contribute

We are actively looking for contributors! Any contributions are welcome, not just code! Please chat to us <https://gitter.im/sktime/community>_ or raise an issue <https://github.com/alan-turing-institute/sktime/issues/new/choose>_ if you're interested.