High-dimensional Statistics

What it is about

You learn methods for the analysis of data where the number of variables is large. After the module you can apply procedures for controlling false discoveries in large-scale multiple testing, use sparse regression methods (e.g., Lasso) and methods for group-specific or low-dimensional structured signals, as well as use regularization approaches for matrix parameters and graph models.

What you will be able to do

• Control of false discoveries in large-scale multiple testing
• Application of sparse regression methods (e.g., Lasso)
• Design of procedures that exploit low-dimensional structure (e.g., group sparsity)
• Extension of procedures to generalized linear models
• Estimation of matrix-valued parameters with regularization
• Handling high-dimensional graph models

What the module consists of

• Lectureintroduces concepts and methods, develops theoretical properties
• Exercisedeepens understanding through exercises and examples

Teaching method

• LectureIntroduction and exemplary presentation of new concepts and theories
• Tutorial group with problem sheetsDeepening of methods, working on detailed examples and theoretical properties