What's the difference between supervised and unsupervised learning, and how are they used in finance?

Machine learning (ML) in finance is a growing topic on the CFA exam. The fundamental distinction is about whether you have a target variable (label) to predict.

Supervised Learning: You have input features (X) and a known output (Y). The algorithm learns the relationship X -> Y from historical data, then predicts Y for new data.

Finance Applications:

• Credit scoring: Predict default probability (Y = default/no default) from borrower characteristics (income, debt ratio, credit history)
• Stock return prediction: Predict next-month return from factors (value, momentum, quality)
• Fraud detection: Classify transactions as fraudulent or legitimate
• Earnings forecasting: Predict quarterly EPS from fundamental and market data

Common Algorithms:

• Linear/logistic regression (simplest, most interpretable)
• Decision trees and random forests
• Support vector machines
• Neural networks (most flexible, least interpretable)

Unsupervised Learning: You only have input features — no target variable. The algorithm finds hidden patterns, groupings, or structure in the data.

Finance Applications:

• Portfolio clustering: Group stocks by return behavior (not just sector) to build truly diversified portfolios
• Regime detection: Identify market regimes (bull/bear/sideways) from price and volatility patterns
• Anomaly detection: Flag unusual trading patterns without pre-defining what 'unusual' means
• Risk factor discovery: Find hidden factors driving asset returns beyond traditional Fama-French factors

Common Algorithms:

• K-means clustering
• Principal Component Analysis (PCA)
• Hierarchical clustering

Loading diagram...

Hybrid Approach — Semi-Supervised: In practice, financial data often has a small amount of labeled data and a large amount of unlabeled data. Semi-supervised methods use both: train on labeled examples and let the unlabeled data improve the model's understanding of the data distribution.

Key Tradeoffs for Analysts:

Practice ML classification questions in our CFA Level II question bank.