University of Illinois Urbana-Champaign

Nonparametric testing in modern statistics: A personal journey

Gao, Hanjia

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/125689

Description

Title
Nonparametric testing in modern statistics: A personal journey
Author(s)
Gao, Hanjia
Issue Date
2024-07-05
Director of Research (if dissertation) or Advisor (if thesis)
Shao, Xiaofeng
Doctoral Committee Chair(s)
Shao, Xiaofeng
Committee Member(s)
  • Yang, Yun
  • Simpson, Douglas
  • Wang, Yuexi
Department of Study
Statistics
Discipline
Statistics
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
  • Nonparametric Testing
  • High-Dimensional Statistics
  • Functional Data
  • Time Series
Abstract
Nonparametric testing is a fundamental branch of statistics and has numerous applications in modern statistics. This thesis is based on four projects in which we study four different nonparametric testing problems. In the first project, we study the two-sample test and aim to test the equality of two high-dimensional distributions. In particular, we propose a novel studentized test statistic based on the maximum mean discrepancy and establish the asymptotic theory of the proposed test in the high dimensional setting. In the second project, we investigate the change point testing problem for vector-valued time series with both temporal and cross-sectional dependence. By integrating the idea of sample splitting and self-normalization, we propose a dimension-agnostic testing method applicable to low-, medium-, and high dimensional settings, and provide the asymptotic properties of the proposed test both under the null and against the local alternatives. In the third project, we focus on functional time series inference and generalize our test proposed in the previous project to the infinite-dimensional setting. In particular, we propose a fully functional approach based on sample splitting and illustrate it for several testing problems. In the fourth project, we consider the two-sample conditional distribution test and propose a new population-level metric that characterizes the discrepancy between two conditional distributions. We also construct a test statistic using generative adversarial networks and introduce a multiplier bootstrap procedure to approximate the critical value. The asymptotic theory is provided under the null and against the local alternatives, and some preliminary simulations are presented to demonstrate the effectiveness of the test.
Graduation Semester
2024-08
Type of Resource
Thesis
Handle URL
https://hdl.handle.net/2142/125689
Copyright and License Information
Copyright 2024 Hanjia Gao

Owning Collections

Graduate Dissertations and Theses at Illinois PRIMARY
Graduate Theses and Dissertations at Illinois