pyhctsa.operations.information.automutual_info_stats¶

pyhctsa.operations.information.automutual_info_stats(y, max_tau=None, est_method='gaussian', extra_param=None)¶

Calculate statistics on the automutual information (AMI) function of a time series.

This function computes various statistics on how the automutual information changes with increasing time delay, including basic statistics, periodicities, and crossings.

Parameters:¶

y : array-like¶

Input time series.

max_tau : int, optional¶

Maximum time delay to investigate. If None, uses N/4 where N is the length of the time series, but won’t exceed N/2. Default is None.

est_method : {'gaussian', 'kraskov1', 'kraskov2'}, optional¶

Method for estimating mutual information (passed to automutual_info). Default is 'gaussian'.

extra_param : int or str, optional¶

Extra parameter for the estimator (passed to automutual_info). For Kraskov estimators, sets the number of nearest neighbors ‘k’. Default is None.

Returns:¶

Dictionary of AMI statistics: the AMI at each lag (ami1 … ami<max_tau>), mami, stdami, pextrema, fmmi (lag of the first local minimum of the AMI function; the number of lags if there is none), sumami_fmmi (the sum of the AMI from lag 1 to fmmi), pmaxima, modeperiodmax, pmodeperiodmax, pminima, modeperiodmin, pmodeperiodmin, pcrossmean, pcrossmedian, pcrossq10, pcrossq90 and amiac1. pmaxima and pminima are the number of intervals between successive local maxima (minima) divided by the integer part of half the number of lags.

Return type:¶

dict