Utils

The documentation of the approximation module.

The pyts.utils module includes utility functions.

pyts.utils.segmentation(ts_size, window_size, overlapping, n_segments=None)

Compute the indices for Piecewise Agrgegate Approximation.

Parameters:

ts_size : int

The size of the time series.

window_size : int

The size of the window.

overlapping : bool

If True, overlapping windows may be used. If False, non-overlapping are used.

n_segments : int or None (default = None)

The number of windows. If None, the number is automatically computed using window_size.

Returns:

start : array

The lower bound for each window.

end : array

The upper bound for each window.

size : int

The size of start.

pyts.utils.numerosity_reduction(arr)

Perform numerosity reduction.

Parameters:

arr : array-like, shape [n_samples]

Returns:

res : str

string with each word separated with a whitespace.

pyts.utils.dtw(x, y, dist=u'absolute', return_path=False, **kwargs)

Dynamic Time Warping.

Parameters:

x : array-like, shape [n1]

First array.

y : array-like, shape [n2]

Second array

dist : {‘absolute’, ‘square’ or callable} (default = ‘absolue’)

The distance metric used. If callable, the first two arguments must be float numbers.

return_path : bool (default = False)

If true, the path along the accumulated cost matrix is returned.

kwargs

Additional keyword arguments for dist if dist is callable. Ignored otherwise.

pyts.utils.fast_dtw(x, y, window_size, approximation=True, dist=u'absolute', return_path=False, **kwargs)

Fast Dynamic Time Warping.

Parameters:

x : array-like, shape [n1]

First array.

y : array-like, shape [n2]

Second array

window_size : int

The size of the window for the PAA algorithm.

approximation : bool (default = True)

If True, compute Dynamic Time Warping between the shrunk time series. If False, compute Dynamic Time Warping on the original time series with a constraint region based on the path of the Dynamic Time Warping of the shrunk time series.

dist : {‘absolute’, ‘square’ or callable} (default = ‘absolue’)

The distance metric used. If callable, the first two arguments must be float numbers.

return_path : bool (default = False)

If true, the path along the accumulated cost matrix is returned.

kwargs

Additional keyword arguments for dist if dist is callable. Ignored otherwise.