Lili Yu
Georgia Southern University, GA·Mathematics & Statistics
How long until a disease appears, a patient recovers, or a patient dies? Survival analysis is the branch of statistics that answers 'time-to-event' questions like these.
- Fields
Mathematics & Statistics
Public Health & Epidemiology
- Topics
survival analysis
time-to-event data
accelerated failure time model
- Grants
- 1 known grant, through Jul 2028
What they work on
- Develop Laplace-approximated quasi-likelihood methods for the semiparametric accelerated failure time (AFT) model, adding higher polynomial terms to reduce bias for data with constant or varying variability.
- Extend the approach to penalized quasi-likelihood methods for AFT models with frailty, a random effect capturing unmeasured shared risk.
- Create regularized penalized quasi-likelihood methods for frailty AFT models that perform variable selection, then build efficient algorithms and open software and test them on real datasets.
More about the research
It shows how treatments or exposures change patient outcomes over time. One flexible tool for this is the semiparametric accelerated failure time (AFT) model. It describes how factors speed up or slow down the time until an event, without assuming a fixed shape for the underlying distribution. Current ways of fitting it either demand heavy computing or rely on rough approximations that can bias results.
Why it matters
Doctors and public health researchers rely on survival analysis to judge how treatments and risk factors affect outcomes such as recovery or death. More accurate, efficient methods with openly available software could give less biased estimates of these effects. That could support better decisions on disease prevention and treatment.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Hong Zhu
University of Virginia, VA·Mathematics & Statistics
Develop a copula-based model, a statistical tool that links the timing of two related events, with time-varying effects and random effects, for clustered semi-competing risks data such as cancer recurrence and death.
Richard Guo
University of Michigan, MI·Mathematics & Statistics
Develop a hunt-and-test version of the classical score test for infinite-dimensional parameters such as regression functions and log-densities, using machine learning to find signals and debiasing to keep the test calibrated.
Peng Liu
Iowa State University, IA·Mathematics & Statistics
Develop functional regression models for imaging outcomes built on a distributed learning framework, so that estimation and inference stay accurate yet scalable for large imaging datasets.
The school
Georgia Southern University: 3 labs on PI Pages, $46M a year on research, 70 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.