Prior distributions for response probability

badr supports PointMass, Beta, BetaMixture, and JeffreysPrior distributions for the reponse probability. The probability density function and cumulative probability functions are available via density and cdf methods. The mean of a distribution can quickly be accessed via mean.

# S4 method for Prior
density(x, p)

cdf(prior, p, ...)

# S4 method for Prior,numeric
cdf(prior, p)

# S4 method for Prior
mean(x)

PointMass(p)

Beta(a, b)

Beta_mu_sd(mu, sd)

JeffreysPrior(design)

Arguments

x	Prior distribution object (density)
p	probability atom, i.e. the response probability equals p almost surely.
prior	Prior distribution object
a	Beta distribution parameter
b	Beta distribution paramter
mu	mean parameter
sd	standard deviation paramter
design	Design object

See also

condition for restricting the support of a prior, updating for Bayesian posterior updates

Examples

badr::load_julia_package()

if (FALSE) {
  density(Beta(1, 1), seq(0, 1, .1)) == 1 # uniform distribution on [0, 1]
}
if (FALSE) {
  cdf(PointMass(1/3), c(0.3, 1/3)) == c(0, 1)
}
if (FALSE) {
  mean(Beta(5, 7)) == 5/(5 + 7)
}
if (FALSE) {
  PointMass(0.4)
}
if (FALSE) {
  Beta(5, 7)
  1/3*Beta(5, 7) + 2/3*Beta(1,1) # create a BetaMixture distribution
}
if (FALSE) {
  Beta_mu_sd(.3, .2)
}
if (FALSE) {
  design <- Design(c(0, 30, 25, 0), c(Inf, 10, 7, -Inf))
  JeffreysPrior(design)
}