Free module · Machine learning & adaptive decisions
Machine Learning for Quantitative Strategy Development
Choose a useful decision, build an honest dataset, and connect model output to net portfolio value.
labels · regression · classification · trees · calibration · representations
The building blocks
A supervised model learns an input-to-output rule from examples. A forecast becomes a trading decision only after you account for costs, uncertainty and portfolio constraints.
- Define one input row and its future target
- Fit a simple baseline on past data
- Validate calibration and the resulting decision
Lessons in this module
- Choose the model’s job: targets, horizons and decision layers
- Feature engineering, missingness and training-only transformations
- Regularised regression: an interpretable alpha baseline
- Logistic classification and cost-aware entry thresholds
- Trees and boosting: nonlinear interactions with controlled complexity
- Calibration, meta-labels and conditional payoff estimation
- Unsupervised learning, clusters and latent risk structure
- From model forecasts to a constrained strategy
Practice and apply
- Build an executable return label — Entry 100.10; exit 100.40.
- Solve one-feature ridge by hand — x=[−1,1], y=[−2,2], penalty λ=2.
Work through the practice exercises · Quant development tools