get_outliers#

pymultifracs.robust.get_outliers(wt_coefs, scaling_ranges, pelt_beta, threshold, pelt_jump=1, robust_cm=False, verbose=False, generalized=False, remove_edges=False)#

Detect outliers in a signal.

Parameters:
wt_coefsWaveletDec

Input coefficients of the signal with outliers.

scaling_rangeslist[tuple[int, int]]

List of pairs of (j1, j2) ranges of scales for the linear regressions.

pelt_betafloat

Regularization parameter for the PELT segmentation.

thresholdfloat

Wasserstein distance threshold to indentify a segment as outlier.

pelt_jumpint

Optional, PELT algorithm checks segmentations every pelt_jump point.

robust_cmbool

Whether to use robust cumulants in the detection.

generalizedbool

Whether to use the exponential power distribution model instead of the normal distribution for the log 1-leaders in the detection.

verbosebool, optional

Display figures outlining the detection process. If multiple signals are being processed, will only show figures for the first signal.

Returns:
leadersWaveletLeader

Wavelet 1-leaders used in the analysis.

idx_rejectdict[int, ndarray]

Dictionary associating to each scale the boolean mask of indices to reject.

See also

mfa()

Can be fed the output dictionary: idx_reject.