find_outliers

sofia_redux.toolkit.stats.stats.find_outliers(data, threshold=5, keepnans=False, axis=None)

Determines the outliers in a distribution of data

Computes the median and the median absolute deviation (MAD) 1.482 * median(abs(x_i-x_med)) of the data and identifies data values as outliers if abs(x_i - x_med) / MAD > threshold where x_med is the median.

Taken from mc_findoutliers in spextool

Parameters:

dataarray_like of (int or float)

thresholdint or float, optional

The sigma threshold

keepnansbool, optional

If True, do not flag NaNs as outliers

axisint, optional

If axis is set, outliers are determined for each slice along the set axis. i.e. if axis=0 for a 2-d array, then outliers will be determined by using statistics derived along each row rather than over the entirety of the data set. This could be is useful when each “row” contains a unique dataset such as a spectrum.

Returns:

numpy.ndarray

boolean array of the same shape as data. False indicates an outlier