tslearn-team/tslearn

The machine learning toolkit for time series analysis in Python

tslearn – Machine‑learning toolkit for time‑series in Python

What it is – A Python library that extends the familiar scikit‑learn API to work with time‑series data. It provides utilities for loading, preprocessing, and generating time‑series, plus a collection of algorithms for classification, clustering, regression, and similarity measurement.

Key capabilities

  • Data handling – Convert lists, other toolkits, or UCR benchmark datasets into the required 3‑D NumPy array (n_ts, max_len, dim). Supports variable‑length series.
  • Pre‑processing – Scaling, resampling, piece‑wise transforms, and synthetic generators.
  • Algorithms
    • Classification: K‑Nearest‑Neighbour, Support‑Vector Classifier, Learning Shapelets, Early Classification.
    • Regression: K‑Nearest‑Neighbour Regressor, Support‑Vector Regressor, MLP.
    • Clustering: TimeSeriesKMeans, K‑Shape, Kernel K‑Means.
    • Metrics: Dynamic Time Warping, Global Alignment Kernel, matrix‑profile, barycenters.
  • Full scikit‑learn compatibility – Models can be used inside pipelines, grid‑search, and other scikit‑learn utilities.

Installation

# PyPI (recommended)
python -m pip install tslearn

# Conda
conda install -c conda-forge tslearn

# From source
python -m pip install https://github.com/tslearn-team/tslearn/archive/main.zip

(Requires Python 3.10+ and the usual scientific‑Python stack.)

Typical workflow

  1. Load/format data → to_time_series_dataset or one of the built‑in loaders.
  2. Pre‑process with TimeSeriesScalerMinMax, resampling, etc.
  3. Fit a model, e.g. KNeighborsTimeSeriesClassifier.
  4. Predict / evaluate using the same scikit‑learn‑style methods.
  5. Optionally compute distances, barycenters, or other analyses.

Documentation & examples – Full API reference, user guide, and a gallery of Jupyter notebooks are hosted on Read the Docs: https://tslearn.readthedocs.io.

Community – Open‑source on GitHub, with contribution guidelines, an issue tracker for feature requests, and a citation ready for academic use (JMLR 2020).

Related

  • Project
    aeon-toolkit/aeon

    aeon is a Python library (scikit‑learn compatible) that provides a unified API for a wide range of time‑series machine‑learning tasks, including state‑of‑the‑art classification, regression, clustering, forecasting, anomaly detection, segmentation, similarity search, transformations, distances/kernels and benchmarking. It bundles both classic statistical models and deep‑learning architectures, is maintained by original algorithm authors, and offers extensive documentation, examples, and community support.

  • Project
    skforecast/skforecast

    A Python library for time series forecasting that integrates scikit-learn compatible models, statistical methods, and foundation models into a unified forecasting workflow.

  • Project
    Nixtla/mlforecast

    A scalable framework for time series forecasting that allows any scikit-learn compatible machine learning model to be used for forecasting millions of series.

  • Project
    sktime/sktime

    A Python library providing a unified interface for time series analysis, including forecasting, classification, and anomaly detection, compatible with the scikit-learn ecosystem.

  • Project
    nilearn/nilearn

    Nilearn is a Python library for the statistical analysis of brain volumes and surfaces, providing GLM-based tools and scikit-learn integration for predictive modeling and classification.