mlots.transformation package¶

Module contents¶

class mlots.transformation.MINIROCKET(num_kernels=10000, max_dilations_per_kernel=32, ts_type='univariate', random_seed=1992)¶

Bases: object

NAME: MINIROCKET

This is a class that represents MINIROCKET by et al [1].

Parameters

num_kernels (int (default 10,000)) – The number of random convolution kernels to be used.
max_dilations_per_kernel (int (default 32)) – The maximum dilation allowed per kernel.
ts_type (str (default "univariate")) – The type of time-series data; either univariate or multivariate.
random_seed (int (default 1992)) – The initial seed to be used by random function.

Returns

object – MINIROCKET class with the parameters supplied.

Return type

self

Raises

ValueError – If the ts_type supplied is incompatible, i.e. not univariate or multivariate.

Examples

>>> from mlots.transformation import MINIROCKET
#ts_type denotes if we are using univariate or multivariate version of the algorithm
#depending upon time-series being univariate or multivariate; choose the ts_type accordingly.
>>> minirocket = MINIROCKET(ts_type="univariate")
>>> minirocket.fit(X_train)
>>> X_train_transformed = minirocket.transform(X_train)
>>> X_test_transformed = minirocket.transform(X_test)

Notes

[1] A. Dempster, D. F. Schmidt, and G. I. Webb. MINIROCKET: A very fast (almost) deterministic transform: for time series classification. arXiv:2012.08791, 2020.

fit(X=None)¶

Parameters: X (ndarray (default None)) – The time-series data to be used to derive the kernels.
Returns: object – MINIROCKET class with the the fitted kernels.
Return type: self

transform(X=None)¶

Parameters: X (ndarray (default None)) – The time-series data to be transformed.
Returns: X – The time-series data after transformation.
Return type: ndarray

class mlots.transformation.ROCKET(num_kernels=10000, random_seed=1992)¶