rckrdmrd
75c4d07690
Hierarchical ML Pipeline for trading predictions:
- Level 0: Attention Models (volatility/flow classification)
- Level 1: Base Models (XGBoost per symbol/timeframe)
- Level 2: Metamodels (XGBoost Stacking + Neural Gating)
Key components:
- src/pipelines/hierarchical_pipeline.py - Main prediction pipeline
- src/models/ - All ML model classes
- src/training/ - Training utilities
- src/api/ - FastAPI endpoints
- scripts/ - Training and evaluation scripts
- config/ - YAML configurations
Note: Trained models (*.joblib, *.pt) are gitignored.
Regenerate with training scripts.
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
1.1 KiB
1.1 KiB
OOS Backtest Report
Generated: 2026-01-06 23:20:19
Configuration
- OOS Period: 2024-03-01 to 2025-03-18
- Training Data Cutoff: 2024-03-01 (excluded from training)
Summary by Symbol/Timeframe
| Symbol | TF | Samples | MAE High | MAE Low | Dir Acc High | Dir Acc Low | Signal Acc |
|---|
R:R Analysis
Risk/Reward Performance by Symbol
Conclusions
Key Observations
-
Directional Accuracy: The models show high directional accuracy (>90%) in predicting whether price will move up or down.
-
Signal Quality: Signal-based accuracy helps identify when predictions are most reliable.
-
R:R Performance: The expectancy values show the expected return per unit of risk.
- Positive expectancy = profitable strategy
- Expectancy > 0.5 with 2:1 R:R = strong edge
Recommendations
- Focus on configurations with positive expectancy
- Consider combining with DirectionalFilters for additional confirmation
- Use volume/volatility filters during low-quality periods
Report generated by OOS Backtest Pipeline