rms: plot.contrast.rms – R documentation

Pricing

Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!

Get Started for Free

Documentation

rms

plot.contrast.rms

Description

Plot Bayesian Contrast Posterior Densities

Usage

## S3 method for class 'contrast.rms'
plot(
  x,
  bivar = FALSE,
  bivarmethod = c("ellipse", "kernel"),
  prob = 0.95,
  which = c("both", "diff", "ind"),
  nrow = NULL,
  ncol = NULL,
  ...
)

Arguments

`x`	the result of `contrast.rms`
`bivar`	set to `TRUE` to plot 2-d posterior density contour
`bivarmethod`	see `rmsb::pdensityContour()`
`prob`	posterior coverage probability for HPD interval or 2-d contour
`which`	applies when plotting the result of `contrast(..., fun=)`, defaulting to showing the posterior density of both estimates plus their difference. Set to `"ind"` to only show the two individual densities or `"diff"` to only show the posterior density for the differences.
`nrow`	for `ggplot2::facet_wrap()`
`ncol`	likewise
`...`	unused

Details

If there are exactly two contrasts and bivar=TRUE plots an elliptical or kernal (based on bivarmethod posterior density contour with probability prob). Otherwise plots a series of posterior densities of contrasts along with HPD intervals, posterior means, and medians. When the result being plotted comes from contrast with fun= specified, both the two individual estimates and their difference are plotted.

Value

ggplot2 object

Author(s)

Frank Harrell

rms

Regression Modeling Strategies

v6.2-0

GPL (>= 2)

Authors

Frank E Harrell Jr <fh@fharrell.com>

Initial release

2021-03-17