mh_mcmc: mh_mcmc

In bcallaway11/qrme: Quantile Regression with Measurement Error

Description

A Metropolis-Hastings algorithm for drawing measurment errors.

Usage

mh_mcmc(
  startval = 0,
  iters = 500,
  burnin = 100,
  drawsd = sqrt(4),
  betmat,
  m,
  pi,
  mu,
  sig,
  y,
  x,
  tau
)

Arguments

startval	The first value in the markov chain
iters	The total number of measurement error draws to make
burnin	The number of draws to drop
drawsd	Trial values are drawn from N(0, sd=drawsd), default is 4
betmat	LxK matrix of parameter values with L the number of quantiles and K the dimension of the covariates
m	The dimension of the measurement error
pi	The probability of each mixture component (should have length equal to m)
mu	The mean of each mixture component (should have length equal to m)
sig	The standard deviation of each mixture component (should have length equal to m)
y	particular value of y
x	particular value of x
tau	an L-vector of all the quantiles where betas were estimated

Value

vector of draws of measurement error

bcallaway11/qrme documentation built on June 30, 2021, 12:52 p.m.