|
Activity Number:
|
322
- Analyses in Ecology, Epidemiology, and Environmental Policy
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 11, 2021 : 3:30 PM to 5:20 PM
|
|
Sponsor:
|
Section on Statistics and the Environment
|
|
Abstract #318068
|
|
|
Title:
|
Autocart: Spatially Aware Regression Trees for Ecological and Spatial Modeling
|
|
Author(s):
|
Ethan Ancell* and Brennan Bean
|
|
Companies:
|
Utah State University and Utah State University
|
|
Keywords:
|
Machine learning;
ecology;
spatial statistics;
regression trees;
autocorrelation
|
|
Abstract:
|
Many ecological and spatial processes are complex in nature and are not accurately characterized by linear models. Machine learning methods promise to handle the high-order interactions that are present in ecological and spatial datasets but often fail to produce physically realistic characterizations of the underlying landscape. The "autocart" (autocorrelated regression trees) R package extends the functionality of existing spatial regression tree methods through a spatially aware splitting function and adaptive inverse distance weighting interpolation in each terminal node. The efficacy of autocart is demonstrated on several spatial datasets. In each case, the autocart approach produces a physically realistic representation of the response variable across geographical space without sacrificing model accuracy. Most importantly, the applications demonstrate the computational feasibility and accessibility of spatially aware regression trees for general use in R.
|
Authors who are presenting talks have a * after their name.
