msmbuilder.msm_analysis.msm_acf

msmbuilder.msm_analysis.msm_acf(tprob, observable, timepoints, num_modes=10)[source]

Calculate an autocorrelation function from an MSM.

Rapid calculation of the autocorrelation of an MSM is performed via an eigenmode decomposition.

Parameters:

tprob : matrix

Transition probability matrix

observable : ndarray, float

Vector representing the observable value for each state

timepoints : ndarray, int

The timepoints at which to calculate the decay, in units of lag times.

num_modes : int (num_modes)

The number of eigenmodes to employ. More modes, more accurate, but slower.
Returns:

acf : ndarray, float

The autocorrelation function.

Notes

Use statsmodels.tsa.stattools.acf if you want to calculate an ACF from a raw observable such as an RMSD trace.

See Docs/ACF/acf.pdf for a derivation of this calculation.

Versions