entroscope¶
The definitive entropy toolkit for time series data. Nine measures, one consistent interface, native pandas/polars/numpy support.
- Quickstart — install and first entropy in 60 seconds.
- Validation — how every measure is checked against antropy, EntropyHub and scipy.
- Integrations — polars input and a scikit-learn transformer.
- Examples — food trends, finance, biomedical, medical & biomedical (HRV, EEG, respiration, glucose), business & operational (sales, web traffic, prices, QC).
Runnable versions of the medical and business examples live in the
examples/
directory — python examples/medical.py and python examples/business.py.
Measures at a glance¶
| Measure | What it captures |
|---|---|
| shannon | Uncertainty in a binned distribution |
| permutation | Ordinal-pattern complexity (robust to noise) |
| sample | Regularity / predictability |
| approximate | Regularity (less noise-sensitive, faster) |
| spectral | Spread of the power spectrum (frequency domain) |
| differential | Continuous entropy via a fitted distribution |
| multiscale | Sample entropy across coarse-grained time scales |
| transfer | Directional information flow X → Y (KSG/binned) |
| divergence | KL and Jensen-Shannon distance between samples |