Model-free knockoffs: high-dimensional variable selection that controls the false discovery rate
Many contemporary large-scale applications involve building interpretable models linking a large set of potential covariates to a response in a nonlinear fashion, such as when the response is binary. Although this modeling problem has been extensively studied, it remains unclear how to effectively control the fraction of false discoveries even in high-dimensional logistic regression, not to mention general high-dimensional nonlinear models. To address such a practical problem, we...
Room number should be posted inside the Countway Library, at the entrance
Replicability Concerns in Medical Research
(Seminar at Harvard Medical School Center for Biomedical Informatics)
I shall describe the recent concerns about increased replicability problems in medical research. I'll review the efforts in the direction of improved reproducibility, and discuss the lack of attention to the problem of selective inference that hampers replicability, as reflected in the recent study of the science-wise FDR level in leading medical journal by Jager and Leek, and its discussions.
I'll present an FDR controlling method that addresses selective inference while attending... Read more about HMS Seminar: Yoav Benjamini