SigmaField
Start learning

Courses Machine Learning for Finance

Course 5Coming soon

Machine Learning for Finance

From linear models to transformers and reinforcement learning on financial data, validated rigorously, with the reasons each model can fail.

  • 16 modules
  • Before you start: Foundations Levels 3, 4 and 5; Financial Econometrics

Coming soon

Explore the syllabus below, and start with the courses it builds on.

How each module works

  1. 1The market problemWhy practitioners need the model
  2. 2Model and codeThe model, then its implementation
  3. 3Practice setQuick questions with instant feedback
  4. 4Case studyAn industry problem, solved in code
  5. 5Limits and risksWhen the model fails, and what to do

Modules

1.1The ML workflow in financeComing soon
You will learn:
Framing, Leakage and look-ahead bias, Splitting in time.
Case study:
Why a "95% accurate" predictor is wrong
1.2Data, features and labelsComing soon
You will learn:
Returns, Alpha factors, Alternative data, Fixed-horizon and triple-barrier labels.
Case study:
A factor dataset for 500 stocks
1.3Linear modelsComing soon
You will learn:
OLS, Ridge, Lasso, Elastic net, Factor models as regression.
Case study:
Shrinkage for return prediction
1.4ClassificationComing soon
You will learn:
Logistic regression, ROC, Precision, Recall, Calibration.
Case study:
Predicting loan default
1.5Validating models on time seriesComing soon
You will learn:
Walk-forward, Purged and embargoed CV, Multiple testing, Deflated Sharpe.
Case study:
Catch an overfit strategy
1.6Trees and ensemblesComing soon
You will learn:
Trees, Random forests, Gradient boosting, SHAP.
Case study:
Cross-sectional stock ranking
1.7Unsupervised learningComing soon
You will learn:
PCA, Clustering, Regime detection, Hierarchical clustering.
Case study:
Detect market regimes
1.8Text and languageComing soon
You will learn:
Bag of words, Sentiment, Embeddings, Transformers.
Case study:
Earnings-call sentiment and returns
1.9Neural network foundationsComing soon
You will learn:
Perceptron to multilayer network, Backpropagation as the chain rule.
Case study:
A network from scratch in NumPy
1.10Deep learning for time seriesComing soon
You will learn:
CNNs, Recurrent networks, Attention, Transformers.
Case study:
Volatility forecasting against GARCH
1.11Latent factor and generative modelsComing soon
You will learn:
Autoencoders, Conditional autoencoder asset pricing, Synthetic data.
Case study:
Latent risk factors from returns
1.12Causal machine learningComing soon
You will learn:
Prediction vs cause, Treatment effects for strategies.
Case study:
Did the signal cause the returns?
1.13Reinforcement learningComing soon
You will learn:
MDPs, Dynamic programming, Q-learning, Policy gradients.
Case study:
Optimal execution and deep hedging
1.14Language models in researchComing soon
You will learn:
Retrieval-augmented research, Agents, Their limits.
Case study:
A filings question-answering tool
1.15From model to strategyComing soon
You will learn:
Signals to portfolios, Costs, Risk, Model governance, MLOps.
Case study:
Production checklist for a model
1.16Capstone: an end-to-end ML strategyComing soon
You will learn:
Data, Features, Model, Honest validation, Costs, Report.
Case study:
Graded with a checklist