In andyhoegh/UncertainOrd: Accounts for Uncertainty in Model-Based Unconstrained Ordination

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Spider Data

As an example, the spider data set is used for ordination. First the abundance data is converted to presence/absence.

#library(mvabund)
#library(knitr)
#library(dplyr)
#install_github('andyhoegh/UncertainOrd')
#library(UncertainOrd)
#library(tidyr)
#library(ggplot2)
data(spider, package='mvabund')
spider.matrix <- matrix(as.numeric(spider$abund > 0), nrow = 28, ncol = 12)
knitr::kable(spider.matrix, col.names = colnames(spider$abund))

Ordination Model

The latent factor model is used for ordination.

ord <- UncertainOrd::ordinate_probit(500, spider.matrix)

Data Visualization

Using the output from the model above, we can create ordination figures.

UncertainOrd::CredibleViz(ord$z.samples[,,1], ord$z.samples[,,2], type = 'points')
UncertainOrd::CredibleViz(ord$z.samples[,,1], ord$z.samples[,,2], type = 'scatter', items = c(1,10,16))
UncertainOrd::CredibleViz(ord$z.samples[,,1], ord$z.samples[,,2], type = 'circles', items = c(1,10,16))
UncertainOrd::CredibleViz(ord$z.samples[,,1], ord$z.samples[,,2], type = 'density', items = c(1,10,16))

andyhoegh/UncertainOrd documentation built on Aug. 9, 2019, 4:33 p.m.