PID Estimators - pwollstadt/IDTxl GitHub Wiki

Different partial information decomposition (PID) estimators for discrete data are available, see the Installation and Requirements page for software requirements.

References

• PID: Williams & Beer, 2010, arXiv:1004.2515 [cs.IT]

Available Estimators

As of May 2020, the following estimators are available:

• SydneyPID (for discrete data)
• TartuPID (for discrete data): Implementation of the BROJA estimator (Bertschinger et al., 2014, Entropy,16(4); Makkeh et al., 2017, Entropy, 19(10); Makkeh et al., 2018, Entropy, 20(271))
• SxPID (for discrete data): Implementation of the shared exclusion estimator proposed in Makkeh et al., 2020, arXiv:2002.03356 [cs.IT]

For a list of available PID estimators and their documentation, enter:

import idtxl.estimators_pid
import idtxl.estimators_multivariate_pid
help(idtxl.estimators_pid)
help(idtxl.estimators_multivariate_pid)