README.md

In rayisaacalan/BASTION: Bayesian Additive Spanning Tree's Input / Output For Non-Parametric Regression

BASTION

Bayesian Additive Spanning Trees Input/Output for Non-parametric regression

The Bayesian Additive Regression Spanning Trees model is a novel (still in development) ensemble model for non-parametric regression intended to be used on data (especially spatial) that lies on a complex or constrained domain, or a space with irregular shape embedded in Euclidean space. Existing ensemble models for non-parametric regression such as Bayesian Additive Regression Trees (BART) or XGBoost are very popular and effective, but these models rely on binary decision tree partition models as their weak learners. These binary decision tree partitions do not respect domain constraints as they only make splits parallel to Euclidean axes. At the core of the BAST model is a novel weak learner; a random spanning tree manifold partition model. The model is based upon 4 possible moves/graph operations which will eventually comprise a tailored backfitting Markov chain Monte Carlo algorithm.

This package implements those 4 possible graph operations as well as some utility functions to make it easier to utilize them.

To install this package, run the following code:

install.packages("devtools")
devtools::install_github("rayisaacalan/BASTION", dependencies = TRUE)

To install this package and view its vignette, run the following code:

install.packages("devtools")
devtools::install_github("rayisaacalan/BASTION", build_vignettes = TRUE, dependencies = TRUE, force = TRUE)
library(BASTION)
vignette("MCMC", package = "BASTION")

To get started with a simple clustered graph, try the following:

set.seed(1)
coords = data.frame(lon = rnorm(50), lat = rnorm(50))
g = constructGraph(coords, 4)
clust_out = constructClusters(g, 5, minclust = 3)
plot(clust_out$spanning_forest,
     layout = as.matrix(coords),
     vertex.color = clust_out$membership,
     edge.arrow.mode = 0)

rayisaacalan/BASTION documentation built on April 27, 2023, 2:06 p.m.