Marginal False Discovery Rates

By Fenguoerbian in False Discovery Rates R

May 16, 2021

Abstract

In this group seminar, I will talk about the concept of Marginal False Discovery Rates(mFDR) based on the paper "Marginal false discovery rates for penalized regression models" by Patrick Breheny.

Date

May 17, 2021

Time

7:00 PM – 8:00 PM

Location

Room 1114, Scientific Research Laboratory Building

Event

Group seminar

The reference paper proposes a new perspective for defining false discovery rates in the content of penalized regression models. The mFDR is easy to compute and can control the number of noise features picked out by the model under some regularity conditions. When there are non-independent correlation structure among the noise features, the author proposes to use permutation method to estimate mFDR.

img

Posted on:
May 16, 2021
Length:
1 minute read, 62 words
Categories:
False Discovery Rates R
Tags:
False Discovery Rates R
See Also:
Evaluate BLRM using `BLRMeval`
Making Go/No-go decision in proof-of-concept trials
Parallel Computing in R, from parallel to foreach and future