pyhmc.integrated_autocorr3

pyhmc.integrated_autocorr3(x)

Estimate the integrated autocorrelation time, \(\tau_{int}\) of a time series.

This method performancs a summation of empirical autocorrelation function, setting the window based on the initial sequence estimator (Geyer 1992), which stops when the sum of two consecutive elements in the empirica autocorrelation function become negative.

Parameters:

x : ndarray, shape=(n_samples, n_dims)

The time series, with time along axis 0.

Returns:

tau_int : ndarray, shape=(n_dims,)

The estimated integrated autocorrelation time of each dimension in x, considered independently.

References

[R6]	Geyer, Charles J. “Practical markov chain monte carlo.” Statistical Science (1992): 473-483.