Download as iCal file

Testing hypotheses on a tree: new error rates and controlling strategies

By Marina Bogomolov
Location Bloomfield 527
Advisor(s):
Academic Program: Please choose
 
Sunday 14 January 2018, 12:00 - 12:30

note special time

In modern statistical challenges we are often presented with a set of families of hypotheses which are organized hierarchically in a tree structure.  

Each family is selected and tested only if all its ancestor hypotheses are rejected. We formulate a general class of error rates addressing selective inference 

on families of hypotheses which are organized in a tree structure, and propose a hierarchical testing procedure with a guaranteed control of such error rates.

Joint work with Yoav Benjamini, Christine Burns Peterson, and Chiara Sabatti