class ChapmanKolmogorovTest

class deeptime.util.validation.ChapmanKolmogorovTest(lagtimes, predictions, predictions_samples, estimates, estimates_samples, observable)

Test results of the Chapman-Kolmogorov test. See ck_test().

Attributes

err_est

Whether the estimated models contain samples

estimates

Estimates at different lagtimes.

estimates_samples

Returns the sampled estimates, i.e., a list of arrays as described in estimates.

has_errors

Whether the prediction contains samples.

lagtimes

Lagtimes at which estimations and predictions were performed.

n_components

predictions

Returns tested model predictions at different lagtimes

predictions_samples

Returns the confidence intervals of the estimates at different lagtimes (if available)

Methods

plot([height, aspect, conf, color, grid, ...])

Shortcut to deeptime.plots.plot_ck_test().

plot(height=2.5, aspect=1.0, conf: float = 0.95, color=None, grid=None, legend=True, xlabel='lagtime (steps)', ylabel='probability', y01=True, sharey=True, **plot_kwargs)

Shortcut to deeptime.plots.plot_ck_test().

property err_est

Whether the estimated models contain samples

property estimates

Estimates at different lagtimes.

Getter:

each row contains the n observables computed at one of the T lagtimes.

Type:

ndarray(T, n, n)

property estimates_samples

Returns the sampled estimates, i.e., a list of arrays as described in estimates. Can be None.

property has_errors

Whether the prediction contains samples.

property lagtimes

Lagtimes at which estimations and predictions were performed.

property predictions

Returns tested model predictions at different lagtimes

Returns:

Y – each row contains the n observables predicted at one of the T lag times by the tested model.

Return type:

ndarray(T, n)

property predictions_samples

Returns the confidence intervals of the estimates at different lagtimes (if available)

If not available, returns None.

Returns:

  • L (ndarray(T, n)) – each row contains the lower confidence bound of n observables computed at one of the T lag times.

  • R (ndarray(T, n)) – each row contains the upper confidence bound of n observables computed at one of the T lag times.