Catalogues#
catalogues provide a unified interface to list, query, and retrieve
collections of sktime objects such as datasets, estimators, metrics, and
cross-validation strategies.
Catalogues can make reproducing and extending benchmarking studies and competitions easier and convenient, by reducing boilerplate code.
A catalogue can be used to:
inspect what items exist in a given collection,
retrieve items by category (e.g., only datasets, only metrics),
obtain either string specifications or fully constructed objects,
describe metadata about the catalogue such as number of items, authorship, and source.
Forecasting Catalogues#
Dummy catalogue of datasets, forecasters, metrics, and cv. |
Classification Catalogues#
Dummy catalogue of datasets, classifiers, metrics, and cv. |