Online Program Home
My Program

Abstract Details

Activity Number: 250 - Topics in Statistical Learning
Type: Contributed
Date/Time: Monday, July 30, 2018 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistical Learning and Data Science
Abstract #328310
Title: Structure Learning for Phylogenetic Tree with Quantitative Characters
Author(s): Chaoyu Yu* and Mathias Drton
Companies: and University of Washington
Keywords: Tree; Structure Learning; Phylogenetics; High-dimensional
Abstract:

We consider the problem of learning the tree topology in phylogenetic analysis of continuous data. The models we treat assume that the evaluation of the continuous characters along the phylogenetic tree follows a Brownian motion process. We propose and compare three methods for recovery of the tree structure. The first is based on independence tests, and the second is based on the size of sample covariances. As a third option, we derive a structural EM algorithm for searching the tree structure and corresponding parameter estimates that maximize the likelihood of the observed data. We discuss asymptotic consistency of our methods and examine the sample size needed for tree recovery under different scenarios.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program