Myra Samuels Memorial Lecture
Modeling Strategies for Bioinformatic Data
Marty Wells
Cornell University
Start Date and Time: Thu, 28 Apr 2011, 4:30 PM
End Date and Time: Thu, 28 Apr 2011, 5:30 PM
Venue: MATH 175
Abstract:
Modern bioinformatic data sets often consist of large numbers of predictors and small samples sizes. This fact has generated a vast literature on statistical methods for the so-called large p small n problem. Examples in biology include gene expression data from micro-arrays and association studies involving a large number of genetic markers with a given phenotype. In this talk I will discuss the use of some modeling and computational strategies for such problems: random effects to induce shrinkage and for model parsimony; mixtures for empirical Bayes prediction and classification; computation via the EM algorithm; and regularization with