Empirical One-Step Conditional Entropy in Infinite Ergodic Systems: Vanishing Entropy Rate, Sparse-Transition Scaling, and Mittag-Leffler Fluctuations
Home / Papers / Empirical One-Step Conditional Entropy in Infinite Ergodic Systems: Vanishing Entropy Rate, Sparse-Transition Scaling, and Mittag-Leffler Fluctuations
Empirical One-Step Conditional Entropy in Infinite Ergodic Systems: Vanishing Entropy Rate, Sparse-Transition Scaling, and Mittag-Leffler Fluctuations
Authors: Ken-ichi Okubo Date: 2026-06-04 Paper ID: openalex:2606.05685
Summary
This paper investigates the behavior of empirical conditional entropy in weakly chaotic systems characterized by infinite invariant measures. The author proves that standard entropy-rate measures vanish for these intermittent maps, necessitating a focus on finite-time information sums to detect sparse transitions. The work establishes a two-term decay estimate for these sums and connects their fluctuations to Mittag-Leffler distributions, offering a refined perspective on predictability in non-ergodic regimes.
Key Contributions
- Proves that the per-step empirical conditional entropy converges to zero in one-dimensional intermittent maps with infinite invariant measures, rendering standard entropy-rate normalization ineffective for quantifying instability in these systems.
- Identifies that the finite-time information sum captures sparse transitions between laminar phases, providing a non-vanishing measure of symbolic complexity despite the vanishing entropy rate.
- Demonstrates that fluctuations in the information sum for weak-chaos regimes follow normalized Mittag-Leffler laws, providing a rigorous statistical description of long-range dependence in these dynamics.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2606.05685
- paper_id
- 2606.05685
- paper_source
- openalex
- domain
- time-series
- tags
-
- architectures
-
- datasets
-
- concept_slugs
-
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-06-07T06:33:18Z