# stats — Statistics functions module.¶

This module contains utility functions used for statistics in ADMIT.

admit.util.stats.mystats(data)[source]

return raw and robust statistics for a distribution

Parameters: data : array The data array for which the statistics is returned. returns: N, mean, std for the raw and robust resp.
admit.util.stats.reducedchisquared(data, model, numpar, noise=None)[source]

Method to compute the reduced chi squared of a fit to data.

Parameters: data : array like The raw data the fit is based on, acceptable data types are numpy array, masked array, and list model : array like The fit to the data, acceptable data types are numpy array, masked array, and list. Must be of the same length as data, no error checking is done dof : int Number of free parameters (degrees of freedom) in the model noise : float The noise/uncertainty of the data Float containing the reduced chi squared value, the closer to 1.0 the better
admit.util.stats.rejecto1(data, f=1.5)[source]

reject outliers from a distribution using a hinges-fences style rejection, using a mean.

Parameters: data : array f : float The factor f, such that only data is retained between mean - f*std and mean + f*std returns: an array with only data between mean - f*std and mean + f*std
admit.util.stats.rejecto2(data, f=1.5)[source]
reject outliers from a distribution
using a hinges-fences style rejection, using a median.
Parameters: data : array f : float The factor f, such that only data is retained between median - f*std and median + f*std returns: an array with only data between median - f*std and median + f*std
admit.util.stats.robust(data, f=1.5)[source]

return a subset of the data with outliers robustly removed data - can be masked