Time series regression

The sktime.regression module contains algorithms and composition tools for time series regression.

All regressors in sktime can be listed using the sktime.registry.all_estimators utility, using estimator_types="regressor", optionally filtered by tags. Valid tags can be listed using sktime.registry.all_tags.

Composition

ComposableTimeSeriesForestRegressor([...])

Time-Series Forest Regressor.

Deep learning

CNNRegressor([n_epochs, batch_size, ...])	Time Series Convolutional Neural Network (CNN), as described in [1].
TapNetRegressor([n_epochs, batch_size, ...])	Time series attentional prototype network (TapNet), as described in [1].

Distance-based

KNeighborsTimeSeriesRegressor([n_neighbors, ...])

KNN Time Series Regressor.

Dummy

DummyRegressor([strategy, constant, quantile])

DummyRegressor makes predictions that ignore the input features.

Interval-based

TimeSeriesForestRegressor([min_interval, ...])

Time series forest regressor.

Kernel-based

RocketRegressor([num_kernels, ...])

Regressor wrapped for the Rocket transformer using RidgeCV regressor.

Base

BaseRegressor()

Abstract base class for time series regressors.

BaseDeepRegressor([batch_size])

Abstract base class for deep learning time series regression.