Aluode/PerceptionLabPortable

1# Authors: The MNE-Python contributors.2# License: BSD-3-Clause3# Copyright the MNE-Python contributors.4 5import numpy as np6from scipy.stats import zscore7 8 9def _find_outliers(X, threshold=3.0, max_iter=2, tail=0):10    """Find outliers based on iterated Z-scoring.11 12    This procedure compares the absolute z-score against the threshold.13    After excluding local outliers, the comparison is repeated until no14    local outlier is present any more.15 16    Parameters17    ----------18    X : np.ndarray of float, shape (n_elemenets,)19        The scores for which to find outliers.20    threshold : float21        The value above which a feature is classified as outlier.22    max_iter : int23        The maximum number of iterations.24    tail : {0, 1, -1}25        Whether to search for outliers on both extremes of the z-scores (0),26        or on just the positive (1) or negative (-1) side.27 28    Returns29    -------30    bad_idx : np.ndarray of int, shape (n_features)31        The outlier indices.32    """33    my_mask = np.zeros(len(X), dtype=bool)34    for _ in range(max_iter):35        X = np.ma.masked_array(X, my_mask)36        if tail == 0:37            this_z = np.abs(zscore(X))38        elif tail == 1:39            this_z = zscore(X)40        elif tail == -1:41            this_z = -zscore(X)42        else:43            raise ValueError(f"Tail parameter {tail} not recognised.")44        local_bad = this_z > threshold45        my_mask = np.max([my_mask, local_bad], 0)46        if not np.any(local_bad):47            break48 49    bad_idx = np.where(my_mask)[0]50    return bad_idx51