Foundations of Applied Machine Learning
This course introduces core machine learning methods, with applications drawn from financial datasets such as cross-sectional equity returns. The emphasis is on understanding how standard machine learning tools behave in high-dimensional, noisy, and time-dependent settings. Rather than surveying many algorithms, the course focuses on a small set of foundational methods: linear models, dimensionality reduction, tree-based methods, and neural networks. A central theme is the challenge of extracting signal from noisy data, and the importance of careful model validation.
By the end of the course, students will:
Understand the principles of supervised learning and model evaluation
Recognize and mitigate overfitting in high-dimensional settings
Apply regularized linear models to panel data
Use dimensionality reduction techniques (e.g., PCA) to uncover structure
Implement and interpret modern tree-based methods (e.g., gradient boosting)
Understand neural networks as tools for learning representations
Evaluate models using appropriate validation techniques for time-dependent data
Online Quarter: Summer 2026
Online Instructor: TBA