Transfer entropy

Back to Entropy combinations

The transfer entropy between the random variables `X`, `Y`, and `w` is defined by

`T(w, X, Y) = H(w, X) + H(X, Y) - H(X) - H(w, X, Y)`

where `H` is the Shannon differential entropy and `w` is the future of `X`. It measures the amount of directed information flow from `X` to `Y`.

References

Measuring Information Transfer,
Thomas Schreiber,
Physical Review Letters, Volume 85, Number 2, 2000.

See also

Differential entropy

Partial transfer entropy

Remark 1.7.8 - Page generated 31.10.2021 07:59.