Applies machine learning selectively to trading systems for regime detection, prediction, and feature engineering with strict baseline comparisons. Use when the user asks for ML models in trading, regime classifiers, return prediction, feature engineering, or guidance on when ML is justified versus simpler approaches.
Use ML only when it clearly outperforms robust non-ML baselines.
Do not use when a simple rules-based baseline has not been defined or validated yet.
## ML Decision Memo
- Problem framing:
- Baseline:
- Candidate model:
- Feature set:
- Validation protocol:
- Production controls:
- Go/No-go recommendation: