High-Dimensional Statistics
Philippe Rigollet's MIT course covers the finite sample analysis of high-dimensional statistical methods, focusing on proof techniques rather than just results. Topics include high-dimensional regression, matrix estimation, and principal component analysis, with attention to optimality guarantees and minimax lower bounds. The course builds from classical statistical theory toward state-of-the-art methods, closing with a look at open research questions in the field. Materials, published through MIT OpenCourseWare, include lecture notes and problem sets suited to students with a solid background in probability and linear algebra. It is aimed at an advanced audience, such as graduate students in statistics, mathematics, or related quantitative fields, who want to understand the theoretical machinery behind modern high-dimensional data analysis rather than just its applications.