BOSSVS#

class pyts.classification.BOSSVS(word_size: int = 4, n_bins: int = 4, window_size: int | float = 10, window_step: int | float = 1, anova: bool = False, drop_sum: bool = False, norm_mean: bool = False, norm_std: bool = False, strategy: str = 'quantile', alphabet: str | ArrayLike | None = None, numerosity_reduction: bool = True, use_idf: bool = True, smooth_idf: bool = False, sublinear_tf: bool = True)[source]#

Bag-of-SFA Symbols in Vector Space.

Each time series is transformed into an histogram using the Bag-of-SFA Symbols (BOSS) algorithm. Then, for each class, the histograms are added up and a tf-idf vector is computed. The predicted class for a new sample is the class giving the highest cosine similarity between its tf vector and the tf-idf vectors of each class.

Parameters:

word_sizeint (default = 4)

Size of each word.

n_binsint (default = 4)

The number of bins to produce. It must be between 2 and 26.

window_sizeint or float (default = 10)

Size of the sliding window. If float, it represents the percentage of the size of each time series and must be between 0 and 1. The window size will be computed as ceil(window_size * n_timestamps).

window_stepint or float (default = 1)

Step of the sliding window. If float, it represents the percentage of the size of each time series and must be between 0 and 1. The window size will be computed as ceil(window_step * n_timestamps).

anovabool (default = False)

If True, the Fourier coefficient selection is done via a one-way ANOVA test. If False, the first Fourier coefficients are selected.

drop_sumbool (default = False)

If True, the first Fourier coefficient (i.e. the sum of the subseries) is dropped. Otherwise, it is kept.

norm_meanbool (default = False)

If True, center each subseries before scaling.

norm_stdbool (default = False)

If True, scale each subseries to unit variance.

strategystr (default = ‘quantile’)

Strategy used to define the widths of the bins:

‘uniform’: All bins in each sample have identical widths
‘quantile’: All bins in each sample have the same number of points
‘normal’: Bin edges are quantiles from a standard normal distribution
‘entropy’: Bin edges are computed using information gain

alphabetNone, ‘ordinal’ or array-like, shape = (n_bins,)

Alphabet to use. If None, the first n_bins letters of the Latin alphabet are used.

numerosity_reductionbool (default = True)

If True, delete sample-wise all but one occurence of back to back identical occurences of the same words.

use_idfbool (default = True)

Enable inverse-document-frequency reweighting.

smooth_idfbool (default = False)

Smooth idf weights by adding one to document frequencies, as if an extra document was seen containing every term in the collection exactly once. Prevents zero divisions.

sublinear_tfbool (default = True)

Apply sublinear tf scaling, i.e. replace tf with 1 + log(tf).

Attributes:

classes_array, shape = (n_classes,): An array of class labels known to the classifier.
idf_array, shape = (n_features,) , or None: The learned idf vector (global term weights) when use_idf=True, None otherwise.
tfidf_array, shape = (n_classes, n_words): Term-document matrix.
vocabulary_dict: A mapping of feature indices to terms.

References

[1]

P. Schäfer, “Scalable Time Series Classification”. Data Mining and Knowledge Discovery, 30(5), 1273-1298 (2016).

Examples

>>> from pyts.classification import BOSSVS
>>> from pyts.datasets import load_gunpoint
>>> X_train, X_test, y_train, y_test = load_gunpoint(return_X_y=True)
>>> clf = BOSSVS(window_size=28)
>>> clf.fit(X_train, y_train)
BOSSVS(...)
>>> clf.score(X_test, y_test)
0.98

Methods

`__init__`([word_size, n_bins, window_size, ...])
`decision_function`(X)	Evaluate the cosine similarity between document-term matrix and X.
`fit`(X, y)	Compute the document-term matrix.
`get_metadata_routing`()	Get metadata routing of this object.
`get_params`([deep])	Get parameters for this estimator.
`predict`(X)	Predict the class labels for the provided data.
`score`(X, y[, sample_weight])	Return the mean accuracy on the given test data and labels.
`set_params`(**params)	Set the parameters of this estimator.
`set_score_request`(*[, sample_weight])	Configure whether metadata should be requested to be passed to the `score` method.

__init__(word_size: int = 4, n_bins: int = 4, window_size: int | float = 10, window_step: int | float = 1, anova: bool = False, drop_sum: bool = False, norm_mean: bool = False, norm_std: bool = False, strategy: str = 'quantile', alphabet: str | ArrayLike | None = None, numerosity_reduction: bool = True, use_idf: bool = True, smooth_idf: bool = False, sublinear_tf: bool = True) → None[source]#

decision_function(X: ArrayLike) → NDArray[float64][source]#

Evaluate the cosine similarity between document-term matrix and X.

Parameters:

Xarray-like, shape (n_samples, n_timestamps): Test samples.

Returns:

Xarray, shape (n_samples, n_classes): Cosine similarity between the document-term matrix and X.

fit(X: ArrayLike, y: ArrayLike) → Self[source]#

Compute the document-term matrix.

Parameters:

Xarray-like, shape = (n_samples, n_timestamps): Training vector.
yarray-like, shape = (n_samples,): Class labels for each data sample.

Returns:

selfobject

predict(X: ArrayLike) → NDArray[Any][source]#

Predict the class labels for the provided data.

Parameters:

Xarray-like, shape = (n_samples, n_timestamps): Test samples.

Returns:

y_predarray, shape = (n_samples,): Class labels for each data sample.

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → BOSSVS#

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.
False: metadata is not requested and the meta-estimator will not pass it to score.
None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.
str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:

sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED: Metadata routing for sample_weight parameter in score.

Returns:

selfobject: The updated object.

Examples using `pyts.classification.BOSSVS`#

Bag-of-SFA Symbols in Vector Space (BOSSVS)

BOSSVS#

Examples using pyts.classification.BOSSVS#

Examples using `pyts.classification.BOSSVS`#