New outcome space `SequentialPairDistances` (based on the "distribution entropy")

JuliaDynamics / ComplexityMeasures.jl

Estimators for probabilities, entropies, and other complexity measures derived from data in the context of nonlinear dynamics and complex systems

MIT License

48 stars 11 forks source link

New outcome space `SequentialPairDistances` (based on the "distribution entropy") #389

Closed kahaaga closed 5 months ago

kahaaga commented 5 months ago

Fixes #383.

Implements:

ImplementsSequentialPairDistances <: CountBasedOutcomeSpace
Implements DistancePairEncoding <: Encoding.
Adds analytical tests for both.
Doc string with examples (including how to compute the "distribution entropy", and reference to literature)

codecov[bot] commented 5 months ago

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Comparison is base (76f37be) 88.63% compared to head (1603e4d) 88.80%.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## main #389 +/- ## ========================================== + Coverage 88.63% 88.80% +0.16% ========================================== Files 80 82 +2 Lines 2358 2393 +35 ========================================== + Hits 2090 2125 +35 Misses 268 268 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

kahaaga commented 5 months ago

DONE:

[x] Removed the iterator
[x] Switched to a generic outer constructor.
[x] Added a convenience function entropy_distribution and added a documentation reference to it.
[x] Added tests for pretty printing.
[x] Removed reference to internal API in example
[x] Skip embedding step if an AbstractStateSpaceSet is given. Doctoring, code and tests and pretty printing updated accordingly.