New techniques have emerged for both predictive and descriptive learning that help us make sense of vast and complex data sets. The particular focus of this course will be on regression and classification methods as tools for facilitating machine learning. This math-light course is only offered remotely via video segments and TAs will host remote weekly office hours using an online platform such as Google Hangout or BlueJeans. Computing will be done in R.
Limited enrollment! Due to the limited space in this course, interested students should enroll as soon as possible.
- Trevor Hastie Professor of Statistics
- Introduction to supervised learning
- Linear and polynomial regression
- Cross-validation and the bootstrap
- Model selection and regularization methods
- Tree-based methods, random forests and boosting
- Support-vector machines
- Nonlinear methods, splines and generalized additive models
- Principal components and clustering
Introductory courses in statistics or probability (STATS60 or equivalent), linear algebra (MATH51 or equivalent), and computer programming (CS105 or equivalent).