Structural Entropy Dynamics with τ(t) Anomaly Prediction — Residual-Driven Attractors + RK45 Engine

:brain: Structural Entropy Dynamics + τ(t) Prediction

This project simulates a dynamic entropy field over (K, D) space using structural attractors, dynamic residual perturbations, and gradient-based flow.

We simulate the time evolution φ(t) = (K(t), D(t)) using RK45 on a 2D entropy landscape:

  • Dynamic entropy surface H(K, D, t)
  • Moving attractor centers
  • Real residuals as external perturbation driver

Then we extract the structural time τ(t) by integrating entropy gradient norms, and use d²τ/dt² to detect anomaly points.

:white_check_mark: Key Results:

  • Final Structural Attractor:
    ( K^* ≈ 9.95,\quad D^* ≈ 69.02 )
  • Final Structure Time:
    ( τ(100) ≈ 81.4 )
  • Detected τ(t) Anomaly Indices:
    [20, 36, 38, 46, 48, 57, 99, 100]
  • Entropy at φ*: ≈ 0.81431
  • Classifier Report:

Structural Bayesian Inference through Entropy Geometry: An IDS-Based Approach