Gibbs Sample for Single Mean and Variance

Gibbs sampler to sample mean and variance of one numeric variable

gibbs_mean(
  y,
  steps = 1000L,
  burnin = 1000L,
  thin = 1L,
  mu_0 = 0,
  sigma2_0 = 1e+06,
  alpha = 0.001,
  beta = 0.001
)

Arguments

y

a vector of values
steps

number of iterations to run Gibbs sampler for
burnin

number of burn-in iterations to discard before proper steps
thin

thinning factor (default 1)
mu_0

prior mean for mu (default 0)
sigma2_0

prior variance for mu (default 1e6)
alpha

prior shape parameter for sigma2 (default 1e-3)
beta

prior scale parameter for sigma2 (default 1e-3)

Details

Assumes conjugate Normal-Inverse-Gamma priors on mean and variance: $$y \sim \mathrm{Normal}(\mu, \sigma^2)$$