Ronghui (Lily) Xu
 |
|
Recent Research Highlights
apply machine learning methods to predict, as well as develop statistical inference, for complex data type such as competing risks of cancer versus non-cancer mortality, in the presence of high dimensional covariates.
causal inference methodology using propensity scores, instrumental variables, principal stratification, mediation or path specific analysis, for complex data type such as in the above, or for rare events in pregnancy safety data.
With the emergence of biomedical big data,
See CV link on left for more information.
Teaching
Useful books and book chapters
High Dimensional Data Analysis in Cancer Research (ed: Li and Xu). Springer 2008.
Explained variation in proportional hazards regression; in: Handbook of Statistics in Clinical Oncology (ed: Crowley and Hoering). CRC Press/Francis&Taylor Group, 2012.
Goodness-of-fit in survival analysis; in: Encyclopedia of Biostatistics (ed: Armitage and Colton, Vol.4). Wiley, 1998
Robustness in proportional hazards regression; in: Handbook of Survival Analysis (ed: Klein et al.). CRC Press/Francis&Taylor Group, 2014.
R Packages on CRAN
phmm: inference under the proportional hazards mixed-effects model.
TimeVTree: Cox model with time-varying coefficients using a tree-based approach.
CoxR2: an information theoretical R-squared type measure under the Cox model.
R2Addhaz: an R2 measure of explained variation under the additive hazards model.
tsriadditive: an instrumental variable approach (2SRI) for survival and competing risks outcomes under the additive hazards model.
survSens: sensitivity analysis with respect to unmeasured confounding for survival and competing risks outcomes.
CompetingRisk: confidence intervals for the cumulative incidence function (CIF) based on proportional cause-specific hazards modeling.
cmprskcoxmsm: inference on risk (CIF) difference and risk ratio under the Cox marginal structure model for competing risks data.
By colleagues I worked with: