2022 Distinguished Lecture Series in Statistical Sciences: Xihong Lin
Hybrid Event (remote/in-person)
Free Event | Registration Required
Happy to announce that after two years of remote DLSS, this year’s event is to be held in person, in Toronto. Remote participation is also available for those unable to attend in person.
Xihong Lin is a Professor of Biostatistics and Coordinating Director of the Program in Quantitative Genomics of Harvard T.H. Chan School of Public Health, and Professor of Statistics at Harvard Faculty of Arts and Sciences. Dr. Lin’s research interests lie in development and application of statistical and computational methods for analysis of high-throughput genetic and genomic data in epidemiological, environmental and clinical studies, for analysis of complex exposure and phenotype data in observational studies, and statistical learning and inference methods for massive data.
Location 1 (Sep 29):
Chestnut Conference Centre
University Room, 3rd Floor
89 Chestnut Street, Toronto, ON M5G 1R1
Location 2 (Sep 30):
University of Toronto
Maple Room (UY9014), 9th Floor
700 University Avenue, Toronto, ON M5G 1Z5
Thursday, September 29, 2022
Chestnut Conference Centre University Room, 3rd Floor, 89 Chestnut Street
- 3:30 - 4:30
- Lessons Learned from the COVID-19 Pandemic: A Statistician’s Reflection
- In this article, I will discuss my experience as a statistician involved in COVID-19 research in multiple capacities in the last two years, especially in the early phase of the pandemic. I will reflect on the challenges and the lessons I have learned in pandemic research regarding data collection and access, epidemic modeling and data analysis of Wuhan and US data, open science and real time dissemination of research findings, implementation science, media and public communication, and partnerships between academia, government, industry and civil society. I will also make several recommendations on preparing for the next stage of the pandemic and for future pandemics.
- 4:30 - 5:30
Friday, September 30, 2022
University of Toronto Maple Room (UY9014), 9th Floor 700 University Avenue
- 3:30 - 4:30
- Ensemble Methods for Testing a Global Null Hypothesis
- Testing a global null is a canonical problem in statistics and has a wide range of applications. In view of the fact of no uniformly most powerful test, prior and/or domain knowledge are commonly used to focus on a certain class of alternatives to improve the testing power, e.g., the class of alternatives in the scenario of the same effect sign or signal sparsity. However, it is generally challenging to develop tests that are particularly powerful against a certain class of alternatives. In this paper, motivated by the success of ensemble learning methods for prediction or classification, we propose an ensemble framework for testing that mimics the spirit of random forests to deal with the challenges. Our ensemble testing framework aggregates a collection of weak base tests to form a final ensemble test that maintains strong and robust power. The key component of the framework is to introduce a certain random procedure in the construction of base tests. We then apply the framework to four problems about global testing in different classes of alternatives arising from Whole Genome Sequencing (WGS) association studies. Specific ensemble tests are proposed for each of these problems, and their theoretical optimality is established in terms of Bahadur efficiency. Extensive simulations are conducted to demonstrate type I error control and power gain of the proposed ensemble tests. In an analysis of the WGS data from the Atherosclerosis Risk in Communities (ARIC) study, the ensemble tests demonstrate substantial and consistent power improvement compared to other existing tests.