Our research involves developing statistical methods and theories for the analysis
of data as commonly arise in randomized controlled trials and observational studies. In
particular, we are concerned with methods dealing in proper ways with informative
censoring, confounding, the curse of dimensionality, multiple testing, and data adaptive
selection of models. Our philosophy is targeted learning, formalized by our recent work
on targeted maximum likelihood learning, and unified loss based learning. This
statistical approach aims to let the data speak for the purpose of answering a particular
scientific question of interest, and provide robust tests of null hypotheses of interest.
We are continuously concerned with bringing these methods into practice and benchmark
them by the practical performance on simulated and real data.
Please note Web site for the new book, Targeted Learning: www.targetedlearningbook.com
Biology & Genetics
Causal Inference
Clinical Epidemiology
Clinical Trials
Computational Biology/Bioinformatics
Epidemiology
HIV
Longitudinal Data Analysis and Time Series
Loss-Based Estimation with Cross-Validation
Multiple Hypothesis Testing
Software
Link
bias.pboot (with Susan Gruber, Kristin Porter, Maya Petersen, and Yue Wang), Susan Gruber (2010)
Statistical Theory and Methods
Survival Analysis
Media Publications
Prediction
Disease Modeling
Design of Experiments and Sample Surveys
Biology
General Biostatistics
Statistical Models
Stochastic Interventions
Computation
Genetics
No subject area