Aeon

When you need to classify time series, detect anomalies, or forecast temporal data, this gives you scikit-learn compatible access to the aeon toolkit. It covers the full range: classification, regression, clustering, forecasting, segmentation, and similarity search with algorithms like ROCKET, DTW, and Inception networks. The docs are thorough about version differences (targets aeon 1.x, which reworked forecasting from 0.x) and honestly flag experimental modules. Particularly useful if you're dealing with sensor data, financial sequences, or any pattern recognition over time and want proven algorithms without rolling your own temporal feature engineering. Needs Python 3.10+ and explicit installation of aeon extras for deep learning models.