Heidrich, Kiraly, & Ray - sktime - python toolbox for time series | PyData Global 2023

Discover SK-Time, a Python library for time series learning and forecasting, providing a toolbox for building, combining, and evaluating models with pre-processing, feature engineering, and visualization tools.

Key takeaways
  • SK-Time is a Python toolbox for time series learning and forecasting.
  • It enables users to build and combine different forecasting models, including univariate and multivariate models.
  • SK-Time supports pre-processing, feature engineering, and model selection, and provides a convenient interface for users to define their own forecasting models.
  • The library also includes tools for benchmarking, grid search, and random search, making it easy to evaluate and compare different models.
  • SK-Time is compatible with popular datasets and can be used with various backend parallelization libraries.
  • The library is highly extensible and allows users to define their own forecasting models and preprocessors.
  • SK-Time includes support for probabilistic forecasting, including predicted intervals and quantiles.
  • The library also includes tools for visualizing forecasting results, making it easy to understand and communicate the results of forecasting models.
  • SK-Time is a flexible and extensible library that can be used for both research and industry applications.
  • It provides a common interface for different forecasting models, making it easy to switch between models and compare results.
  • The library includes support for multiple languages, including Python, R, and Julia.
  • SK-Time is actively maintained and updated to keep up with the latest developments in the field of time series forecasting.
  • The library is widely used in academia and industry, and has been applied to a wide range of applications, including finance, healthcare, and energy.