Statistics 8932, Spring 2016

Class hours:

Instructor:

Yuhong Yang, 376 Ford Hall

Email: yyang@stat.umn.edu, Phone: 612-626-8337

Office hours:

MW 1:30-3:00

Course description: Nonparametric Function Estimation and Model Selection

Nonparametric methods are important tools for function estimation. We will study the most popular nonparametric function estimation methods and examine fundamental issues, including rate of convergence, trade-off between approximation and estimation, and adaptive estimation. Since a number of parametric and nonparametric methods are available, model selection becomes a crucially important issue in almost all statistical applications. We will cover up-to-date model selection/combining methods and their properties.

Course requirement:

• Homework: a few assignments will be given during the semester.
• There will be no exam. Instead, a project of data analysis and prediction will be done. Also, each student needs to make a presentation on a research topic related to nonparametric function estimation, model selection or model combining. A list of topics and papers will be provided later.

Grading:

• Homework: 30%
• Project: 30%
• Final presentation: 25%
• In-class participation: 15%