infomeasure

Illustrative examples

Small, fully reproducible systems whose information-theoretic behaviour is familiar from the literature. These are illustrative rather than validation, and the parity suite covers accuracy. Every curve is computed with infomeasure and drawn here as inline SVG.

Logistic map, permutation entropy

The logistic map xt+1 = r xt(1 − xt) is the textbook route to chaos. Its order-3 permutation entropy drops at the periodic windows and rises in the chaotic bands. The two estimators have different offsets, so both curves are min-max normalised over the parameter range to make their shapes comparable.

-0.060.220.50.781.13.53.63.73.83.94logistic parameter rentropy (normalised)Permutation entropyGaussian kernel
logistic map, computed with infomeasure

Kuramoto oscillators, MI and TE

A population of phase oscillators with heterogeneous frequencies synchronises as the coupling grows. Mutual information and transfer entropy between two oscillators rise through the transition and the estimators agree in the ordered phase.

-0.3911.43.35.170123456coupling Kvalue (nats)MI, KSGMI, Gaussian kernelTE, KSG
Kuramoto oscillators, computed with infomeasure-rs

Coupled Rossler systems, TE, Renyi and Tsallis

Two unidirectionally coupled Rossler systems are a standard test for causal inference. The signed difference of forward and backward transfer entropy is positive when the coupling direction is recovered. The parametric Renyi and Tsallis families expose a dependence on their parameters that the Shannon estimators cannot.

-0.03870.00440.04750.09060.13400.050.10.150.2coupling edelta TE (nats)KSGGaussian kernelOrdinal
directionality versus coupling, computed with infomeasure-rs
-0.596-0.4-0.203-0.00660.190.611.41.82Renyi order alpha / Tsallis qdelta TE (nats)Renyi, alphaTsallis, q
parametric transfer entropies, computed with infomeasure-rs