Expand description
Context-Aware Regression Predictor (PMAT-039)
Context-aware regression thresholds accounting for system state and historical trends.
§Features
- Context capture (temperature, memory, frequency)
- Adaptive threshold computation based on context
- Trend detection from historical data
- False positive reduction through learned patterns
§Falsification Criteria (F1311-F1320)
See tests/context_regression_f1311.rs for falsification tests.
Structs§
- Baseline
Entry - Historical baseline entry
- Context
Regression Predictor - Context-aware regression predictor
- Regression
Check - Result of regression check
- Regression
Threshold - Computed regression threshold
- System
Context - System context snapshot
- Trend
- Detected trend
Constants§
- DEFAULT_
COLD_ START_ MARGIN - Default cold start margin (15%)
- DEFAULT_
STALENESS_ SEC - Default context staleness (seconds)
- MIN_
SAMPLES_ FOR_ CONTEXT - Default minimum samples for learning