Chen Zehua

Chen Zehua

Professor of Statistics
Department of Statistics and Data Science
National University of Singapore
3 Science Drive 2, Singapore 117543
Republic of Singapore

stachenz@nus.edu.sg

(65) 65166307 (office)

(65) 84995559 (mobile)

(65) 68723919 (fax)

Education

B.S., Wuhan Univ. (1981)

M.S., Univ. of Iowa (1985)

Ph.D., Univ. of Wisconsin-Madison (1989)

Current Research Interests

Model selection criteria.

Feature selection with high dimensional space.

Statistical genetics.

Publications

Publication List

Recent Preprints

o Sequential Lasso cum EBIC for feature selection with ultra-high dimensional feature space (Journal Link) (pdf)

o Extended BIC for linear regression models with diverging number of relevant features and high or ultra-high feature spaces. (Journal Link)

o A two-stage penalized logistic regression approach to case-control genome-wide association studies. Supplementary document (Journal Link)

o Selection Consistency of EBIC for GLIM with Non-canonical Links and Diverging Number of Parameters. Supplementary document (arXiv link)

Monograph

Ranked Set Sampling: Theory and Applications

Z. Chen, Z. D. Bai and B. K. Sinha

Springer ©2004

Table of Contents

Statistical Methods for QTL Mapping

Z. Chen

Chapman & Hall/CRC © 2014

Selected Pages

Teaching Corner

FMS1203S: Randomness in scientific thinking

Course Outline

The purpose of this seminar is to introduce students to the roles of random-
ness in scientic thinking. Some of the topics covered include the following:

• Is probability intuitive? A class exercise will be conducted where stu-
  dents are asked to generate sequences of real and fake random coin
  tosses and are asked to develop tests to detect the dierence.
• What is the role of randomization in the design of scientic exper-
  iments (for instances, why are patients randomly assigned to treat-
  ments in a medical trial)? We recreate a famous incident in which a
  tea time conversation led to a statistician conducting an experiment to
  test whether someone could distinguish whether milk had been added
  rst or last to a cup of tea.
• How has statistical thinking been used and abused in the history of
  IQ testing?
• In the analysis of environmental problems like global warming scientic
  models are often used which are deterministic (roughly speaking, such
  models predict a denite output for a given input). A statistical model
  on the other hand gives predictions in the form of probabilities of
  dierent possible outcomes. How can the deep physical understanding
  embedded in the deterministic models be reconciled with statistical
  approaches to quantifying uncertainty and risk, and why is quantifying
  uncertainty important?
• How can fake random numbers generated on a computer by non-
  random rules sometimes do complicated calculations that aren’t easily
  done by other means?

Instruction Notes

Week 1 Introduction.

Week 2 Problem with sampling. Group 1 Group 4

Week 3 Experimental design. Group 1 Group 2 Group 3 Group 4 Group 5

Week 4 Hypothesis testing. Group 1 Group 2 Group 3 Group 4 Group 5

Week 5 Misleading with or without data Group 1 Group 2 Group 3 Group 4 Group 5

Week 6 Statistics and Environment. Group 1 Group 2 Group 3 Group 4 Group 5

Week 8 Misleading reports in the Media. Group 1 Group 2 Group 3 Group 4 Group 5

Week 9 Reading assignment. Group 1 Group 2 Group 3 Group 4 Group 5

Week 10 Reading assignment. Group 1 Group 2 Group 3 Group 4 Group 5

Week 11 Reading assignment. Group 1 Group 2 Group 3 Group 4 Group 5

Week 12 Reading assignment. Group 1 Group 2 Group 3 Group 4 Group 5

Week 13 Instruction of Final report. Group 1 Group 2 Group 3 Group 4 Group 5

FMS1204S: Fraud, deception and data

Course Outline

The purpose of this seminar is to explore the relationship
between fraud and deception and statistics. Very often
misleading claims arise from an ignorance of basic statistical
ideas, but statistical methods can also be abused knowingly in
fraudulent behavior. On the other hand, statistical methods are
also commonly used to detect and uncover fraud and
dishonesty. This seminar will discuss the role of statistics in
uncovering deception in areas such as:

• Misleading claims in health;
• Misleading surveys and opinion polls;
• How has statistical thinking been used and abused in the history of
  IQ testing?
• Claims and counterclaims in environmental science;
• Fraud detection in the financial world;
• Authorship disputes and detecting plagiarism.

Instruction Notes

Week 1 Introduction.

Week 2 Types of medical data and misinterpretation.

Week 4 Surveys and opinion polls.

Week 5 Data and consumer: danger and opportunity. Group 1 Group 5 Group 2 Group 4

Week 6 Uncertainty and controversy in environmental research. Group 1 Group 2 Group 3 Group 4 Group 5

Week 8 Fraud in the financial world. Group 1 Group 2 Group 3 Group 4 Group 5

Week 9 Fraud in archaeology. Group 1 Group 2 Group 3 Group 5

Week 10 Reading assignment Group 1 Group 2 Group 3 Group 4 Group 5

Week 11 Instruction. Group 2 Group 3 Group 5

Week 12 Instruction. Group 1 Group 2 Group 3 Group 5

Week 13 Instruction. Group 1 Group 3 Group 5

Back to the home page of the Department of Statistics and Applied Probability