Loading repository data…
Loading repository data…
DIMProductions / repository
A lightweight toolkit for auditing non-equilibrium transitions and structural collapse in complex time-series data. Features dual-branch DSP (LF/HF), interpretable metrics, and an interactive browser UI.
A transparent discovery signal based on current public GitHub metadata.
This score does not audit code, security, maintainers, documentation quality, or suitability. Verify the repository and its current documentation before adoption.
Time-Series Event Auditing for Structural Collapse Detection
non-equilibrium-auditor is an experimental toolkit designed to detect and audit non-equilibrium transitions and structural collapse events in complex, noisy multi-channel time-series data.
Unlike traditional anomaly detection, this system focuses on the regime break—the moment a system crosses a stability barrier and enters a collapsed state.
Many real-world systems remain in a stable basin until a perturbation pushes them across a barrier. This transition is rarely a simple spike; it is a structural shift characterized by:
The pipeline separates signal analysis into a Dual Branch architecture to isolate the event from its environment.
Tracks slow environmental behavior, baseline drift, and long-term oscillations. It acts as a reference map to ensure environmental noise isn't mislabeled as a structural event.
Searches for rapid structural changes. This branch evaluates potential candidates using metrics like local SNR, irreversibility, and persistence.
Detected candidates are categorized into a discussable and inspectable hierarchy:
Each event is audited using interpretable physical metrics:
hit_windowsRatio_MirrorSym) often produced by sensor glitches or DSP side-effects.death_audit_v15.py)The core backend that processes raw multi-channel data and ranks events.
index.html)A browser-based analysis dashboard to visualize, verify, and audit detected events without additional software.
© DIMProductions Research use only.