predict.micore: Get predicted OTU abundances

In kevinmcgregor/micore: Microbiome Covariance Regression

Description

Get predicted OTU abundances

Usage

## S3 method for class 'micore'
predict(
  object,
  newdata = NULL,
  type = c("alr", "proportion"),
  post.stat = c("mean", "median"),
  quant = c(0.025, 0.975),
  ...
)

Arguments

object	An object of class micore
newdata	An optional numeric matrix containing covariates for new observations to get predictions for.
type	Character specifying what scale to get predicted OTU abundances for. Additive log-ratios or proportions?
post.stat	Character specifying whether the predictions be based on the posterior mean or median.
quant	Numeric vector specifying the quantiles of the posterior to return for the OTU abundances.
...	Further arguments passed to or from other methods

Value

A list containing:

• fit: A matrix containing the posterior mean (or median) of the OTU abundances on the appropriate scale

• quant: A list containing the requested quantiles from the posterior mean for the OTU abundances.

Examples

n <- 50
p <- 5
X <- cbind(1, rnorm(n))
counts <- matrix(0, n, p+1)
for (i in 1:n) {
  counts[i,] <- rmultinom(1, size=100, prob=rep(1,p+1))
}

library(micore)
mc.fit <- micore(counts, X, n.samp=100, n.burn=100, n.chain=1)

new.dat <- cbind(c(1,1,1),c(0,1,0))
pred <- predict(mc.fit, newdata = new.dat)

pred.p <- predict(mc.fit, newdata=new.dat, "prop")

kevinmcgregor/micore documentation built on June 9, 2021, 10:29 p.m.