SDS Seminar Series – Huiyan Sang, Texas A&M University

Oct

2024

2:00 pm – 3:00 pm

In Person (view details)

Featured Speaker(s): Huiyan Sang

Cost: Free

GS-BART: Graph Split Additive Decision Trees for Spatial and Network Data

Description

The Fall 2024 SDS Seminar Series continues on October 4th from 2:00 p.m. to 3:00 p.m. with Dr. Huiyan Sang (Professor, Department of Statistics, Texas A&M University). This event is in-person in CBA 4.348.

Title: GS-BART: Graph Split Additive Decision Trees for Spatial and Network Data

Abstract: Ensemble decision tree methods such as XGBoost, RF, and BART have gained enormous popularity in data science for their superior performance in machine learning regression and classification tasks. In this paper, we develop a new Bayesian graph-split-based additive decision trees method, called GS-BART, to improve the performance of Bayesian additive decision trees for complex dependent data with graph relations. The new method adopts a highly flexible split rule complying with graph structure to relax the axis-parallel split rule assumption in most existing ensemble decision tree models. We design a scalable informed MCMC algorithm leveraging a gradient-based recursive algorithm on spanning trees or chains to sample the graph-split-based decision tree. The superior performance of the method over conventional ensemble tree models and Gaussian process models is illustrated in various regression and classification tasks for spatial and network data analysis.