pp_contrast: Posterior predictive contrast

In iamamutt/rbaes: R Bayesian Analysis and Estimation Stuff

Description

Posterior predictive contrast

Usage

pp_contrast(stanreg, cdata, ccoef, width = 0.95, yfun = median, ...)

Arguments

stanreg	stanreg object from rstanarm.
cdata	a list of data sets to use to find posterior predictions of each.
ccoef	contrast coefficients corresponding to the order of the data sets in cdata.
width	confidence interval width. Sent to the prob argument from posterior_interval.
yfun	the stat function to use to collapse the predictions into a scalar value. Corresponds to the average of the predicted responses for each data set in cdata. Default is median.
...	additional parameters passed to the function posterior_predict posterior_interval.

Value

list containing the confidence interval and posterior predictive contrast difference

Examples

stanreg <- example_stanreg()
cdata <- list(
c1 = contrast_data(stanreg, TRUE, margin_ignore = 'floor', subset_expression = .~ log_uranium < -0.5),
c2 = contrast_data(stanreg, TRUE, margin_ignore = 'floor', subset_expression = .~ log_uranium >= -0.5 & log_uranium <= 0.5),
c3 = contrast_data(stanreg, TRUE, margin_ignore = 'floor', subset_expression = .~ log_uranium > 0.5))

my_contrast <- pp_contrast(stanreg, cdata, c(-1, 0.5, 0.5))
my_contrast$ci

iamamutt/rbaes documentation built on May 18, 2019, 1:27 a.m.