sktime.transformations.panel.rocket

class sktime.transformations.panel.rocket.MiniRocket(num_features=10000, max_dilations_per_kernel=32, random_state=None)[source]

Bases: sktime.transformations.base._PanelToTabularTransformer

MINIROCKET

MINImally RandOm Convolutional KErnel Transform

Univariate

Unviariate input only. Use class MiniRocketMultivariate for multivariate input.

@article{dempster_etal_2020,

author = {Dempster, Angus and Schmidt, Daniel F and Webb, Geoffrey I}, title = {{MINIROCKET}: A Very Fast (Almost) Deterministic Transform for

Time Series Classification},

year = {2020}, journal = {arXiv:2012.08791}

}

Parameters
  • num_features (int, number of features (default 10,000)) –

  • max_dilations_per_kernel (int, maximum number of dilations per kernel (default 32)) –

  • random_state (int, random seed (optional, default None)) –

fit(X, y=None)[source]

Fits dilations and biases to input time series.

Parameters
  • X (pandas DataFrame, input time series (sktime format)) –

  • y (array_like, target values (optional, ignored as irrelevant)) –

Returns

Return type

self

transform(X, y=None)[source]

Transforms input time series.

Parameters
  • X (pandas DataFrame, input time series (sktime format)) –

  • y (array_like, target values (optional, ignored as irrelevant)) –

Returns

Return type

pandas DataFrame, transformed features

class sktime.transformations.panel.rocket.MiniRocketMultivariate(num_features=10000, max_dilations_per_kernel=32, random_state=None)[source]

Bases: sktime.transformations.base._PanelToTabularTransformer

MINIROCKET (Multivariate)

MINImally RandOm Convolutional KErnel Transform

Multivariate

A provisional and naive extension of MINIROCKET to multivariate input. Use class MiniRocket for univariate input.

@article{dempster_etal_2020,

author = {Dempster, Angus and Schmidt, Daniel F and Webb, Geoffrey I}, title = {{MINIROCKET}: A Very Fast (Almost) Deterministic Transform for

Time Series Classification},

year = {2020}, journal = {arXiv:2012.08791}

}

Parameters
  • num_features (int, number of features (default 10,000)) –

  • max_dilations_per_kernel (int, maximum number of dilations per kernel (default 32)) –

  • random_state (int, random seed (optional, default None)) –

fit(X, y=None)[source]

Fits dilations and biases to input time series.

Parameters
  • X (pandas DataFrame, input time series (sktime format)) –

  • y (array_like, target values (optional, ignored as irrelevant)) –

Returns

Return type

self

transform(X, y=None)[source]

Transforms input time series.

Parameters
  • X (pandas DataFrame, input time series (sktime format)) –

  • y (array_like, target values (optional, ignored as irrelevant)) –

Returns

Return type

pandas DataFrame, transformed features

class sktime.transformations.panel.rocket.Rocket(num_kernels=10000, normalise=True, random_state=None)[source]

Bases: sktime.transformations.base._PanelToTabularTransformer

ROCKET

RandOm Convolutional KErnel Transform

@article{dempster_etal_2019,

author = {Dempster, Angus and Petitjean, Francois and Webb, Geoffrey I}, title = {ROCKET: Exceptionally fast and accurate time series classification using random convolutional kernels}, year = {2019}, journal = {arXiv:1910.13051}

}

Parameters
  • num_kernels (int, number of random convolutional kernels (default 10,000)) –

  • normalise (boolean, whether or not to normalise the input time) –

  • per instance (default True) (series) –

  • random_state (int (ignored unless int due to compatability with Numba),) –

  • seed (optional (random) –

  • None) (default) –

fit(X, y=None)[source]

Infers time series length and number of channels / dimensions ( for multivariate time series) from input pandas DataFrame, and generates random kernels.

Parameters
  • X (pandas DataFrame, input time series (sktime format)) –

  • y (array_like, target values (optional, ignored as irrelevant)) –

Returns

Return type

self

transform(X, y=None)[source]

Transforms input time series using random convolutional kernels.

Parameters
  • X (pandas DataFrame, input time series (sktime format)) –

  • y (array_like, target values (optional, ignored as irrelevant)) –

Returns

Return type

pandas DataFrame, transformed features