utils.parameters.tau_search

utils.parameters.tau_search(series, num_lags=30, return_MI=False)

Find optimal tau value for time delay embedding.

First minimum of mutual information between series and time lagged copies of itself is “optimal” in this case in accordance with Abarnabel’s “Analysis of Observed Chaotic Data”

Parameters

series : pyleo.Series

Series for which we’d like to find the optimal tau value

num_lags : int = 30

Number of time delays to consider. Default is 30

return_MI : (bool, {True, False}) = False

Whether or not to return the list of mutual information values. Useful if the first minimum seems spurious and you’d like to inspect the results.

Returns

tau : int

Optimal time delay parameter according to first minimum of mutual information

MI : list

List of mutual information values. Indices + 1 correspond to amount of lag (index 0 is lag 1, index 1 is lag 2, etc.). Only returned if return_MI is set to True.

Citations

I., Abarbanel Henry D. Analysis of Observed Chaotic Data. Springer, 1997.