Statistics Seminar - Fall 2022

For questions about the seminar schedule, please contact Zuofeng Shang

September 15

Haipeng Xing, Stony Brook University University

Statistical Surveillance of Structural Breaks in Credit Rating Dynamics

Financial crisis usually has severe consequences on the global economy and an intriguing question is whether structural breaks in the credit market can be modeled and surveillanced. Choosing firms’ credit rating transition dynamics as a proxy of the credit market, we approach the problem in two steps. The first step is to model credit rating transitions as a piecewise homogeneous Markov chains with unobserved structural breaks and develop an inference procedure, and the second step is to discuss how statistical process control rules can be used to surveil structural breaks in firms’ rating transition dynamics. Several surveillance rules, such as the LR/GLR rules, the extended Shiryaev’s detection rule, and a Bayesian detection rule for piecewise homogeneous Markovian models. The effectiveness of these rules was analyzed on the basis of Monte Carlo simulations. We also provide a real example that used the surveillance rules to analyze and detect structural breaks in the monthly credit rating migration of U.S. firms from January 1986 to February 2017.

December 1

Yuan Huang, Yale School of Public Health

Subgroups and Multiple Changepoints Detection of Natural Disease History for Huntington’s Disease

Huntington’s Disease (HD) is an inherited neurological disorder caused by a single gene mutation. Spanning a long life-course that usually strikes in early life and manifests in the mid age, HD causes gradual deterioration of cognitive function and motor function that ultimately leads to loss of functional capacity. Despite seemingly straightforward etiology, there is no cure nor an effective treatment. Major challenges in designing effective trials include enormous patient heterogeneity and unclear timing of intervention. To this end, we propose a two-stage regression-based method which simultaneously identifies changepoints and subgroups of unbalanced longitudinal observations. The proposed method advances from the piecewise linear growth mixture model (PLGMM), which 
requires some prior information and often assumes a balanced design. In the proposed method, the realization of simultaneous detection is via the minimax concave penalty (MCP) of adjacent derivative difference and pairwise distance of functions represented by the second order B-spline basis functions. The alternating direction method of multipliers (ADMM) algorithm is developed to obtain the estimated coefficients and the locations of interior knots are fine-tuned to improve accuracy of the changepoints. Compared to the existing methods, the proposed
method shows outstanding performance in numerous simulation studies. We also demonstrate the findings of our method on the Enroll-HD dataset. 

December 15

Chong Jin, NJIT

Integrating Multi-Omics Summary Data Using a Mendelian Randomization Framework

Using Mendelian randomization, we can identify the possible causal relationship between an omics biomarker and disease outcome using genetic variants as instrumental variables. This allows us to prioritize genes whose omics readouts can be used as predictors of the disease outcome through analyzing GWAS and QTL summary data. However, best practices are elusive when jointly analyzing the effects of multiple -omics biomarkers annotated to the same gene of interest. To bridge this gap, we propose powerful combination tests that integrate multiple correlated p-values without knowing the dependence structure between the exposures. Our simulation experiments demonstrate the superiority of our proposed approach compared with existing methods adapted to the setting of our interest.

November 23, 2022