Econ 557a Econometrics of High-Dimensional Models

Day / time: 
Tues/Thurs 2:30-3:50
Course Type: 
Graduate
Course term: 
Fall
Visiting Instructor(s): 
Denis Chetverikov
Location: 
Room 108

This class provides an introduction of econometrics of high-dimensional models.  The class will cover the following topics:  1)  relevant results in probability theory (concentration and maximal inequalities); 2) estimation of linear high-dimensional models using Lasso, Dantzig selector, and related methods; 3) estimation of generalized linear high-dimensional models (e.g., quantile and logit regressions) using penalized M-estimators;  4) basics of machine learning (regression trees, random forests, neural networks); 5) semi-parametric inference in high-dimensional models via double machine learning; 6) related topics in econometrics such as grouped fixed effect estimators in panel data and many moment inequalities.  Although the class will be primarily based on research papers, as a general reference, a highly recommended textbook is Wainwright (2019), High-dimensional statistics:  a non-asympotic view point, Cambridge University Press.

Semester offered: 
Fall
Course Description: 
Course Description