CRT approach to anomaly detection. Decompose every measurement into 7 algebraic channels. Normal behavior = consistent channel patterns. mod-11 = instant anomaly flag. No training. No baseline. No model. Demonstrated on synthetic time series with injected anomalies.
Random seed:
Generates 50 data points with periodic signal + injected anomalies. CRT decomposes each, detects via mod-11-weighted channel deviation. Green rows = true positive. Orange = false positive.
20 independent streams, 50 points each. Aggregated precision/recall across all 1000 measurements.
Each anomaly has a unique CRT fingerprint. Per-channel deviations classify the anomaly type automatically.
Source code · Public domain (CC0)
.ax source compiled to WASM via self-hosting compiler. Zero HTML authored.