Tier 4 · QuantFree
Machine Learning in Finance
The Alpha Machine
10 modules~73 min totalVerifiable certificate on completion
Syllabus
01The Bias-Variance TradeoffMath
8 min02Linear Regression and Regularization (Ridge & Lasso)Math
8 min03Cross-Validation and OverfittingMath
9 min04Decision Trees and Gradient BoostingMath
9 min05Feature Importance and the Deflated Sharpe RatioMath
9 min06Overfitting and Backtest Inflation
6 min07Feature Engineering for Alpha
6 min08Tree Models for Alpha Signals
6 min09Sequential Models and Signal Decay
6 min10Alpha Decay and Adversarial Markets
6 minFrom Module 1 — read a sample
Overfitting is memorizing the noise in your training data instead of learning real patterns — like memorizing past exam questions without understanding the concepts, so you fail on new questions. Bias is systematic error (the model is always wrong in the same direction), and variance is instability (retrain on different data, get wildly different answers). The art of machine learning for alpha is balancing these two sources of error.
Teaching a class?
Assign this course as homework. Students sign up free, work through the modules at their own pace, and earn a certificate with a public verification link they submit to you — no teacher account or setup required.
See the educator guide →