aeon

Time-series classification uses algorithms like ROCKET and InceptionTime to identify patterns in sequential data. Aeon provides scikit-learn compatible classifiers that train on temporal sequences, enabling you to apply standard ML workflows to time-series problems without custom preprocessing pipelines.

Aeon includes dedicated forecasters and anomaly detectors designed for long temporal sequences. These components apply to univariate and multivariate data, letting you predict future values and flag deviations from expected patterns in continuous time-series streams.

Aeon provides feature extraction, distance metrics, and transformation utilities tailored for multivariate temporal data. These tools extract discriminative subsequences and transform raw sequences into representations compatible with classification, clustering, and regression tasks.

Aeon includes clustering and segmentation algorithms for grouping similar time series and partitioning sequences into meaningful regions. These components work with the scikit-learn API, integrating seamlessly into end-to-end pipelines for temporal data analysis.

Aeon's scikit-learn compatibility enables you to construct pipelines that chain transformers and estimators, supporting classification, regression, forecasting, and anomaly detection. This allows reproducible, modular workflows for time-series machine learning at scale.