HSI
From: Bayesian Models for Astrophysical Data, Cambridge Univ. Press
(c) 2017, Joseph M. Hilbe, Rafael S. de Souza and Emille E. O. Ishida
you are kindly asked to include the complete citation if you used this material in a publication
​
​
Code 6.18 Synthetic data for generalized Poisson
====================================================
require(MASS)
require(R2jags)
​
source("https://raw.githubusercontent.com/astrobayes/BMAD/master/auxiliar_functions/rgp.R")
set.seed(160)
nobs <- 1000
​
x1 <- runif(nobs)
xb <- 1 + 3.5*x1
exb <- exp(xb)
delta <- -0.3
gpy <- c()
​
for (i in 1:nobs){
gpy[i] <- rgp(1, mu=(1-delta)*exb[i], delta = delta)
}
​
gpdata <- data.frame(gpy, x1)
====================================================
​
​
​
Code 6.19 Bayesian generalized Poisson using JAGS
====================================================
X <- model.matrix(~ x1, data = gpdata)
K <- ncol(X)
​
model.data <- list( Y = gpdata$gpy, # response
X = X, # covariates
N = nrow(gpdata), # sample size
K = K, # number of betas
Zeros = rep(0, nrow(gpdata)))
​
sink("GP1reg.txt")
​
cat("
model{
# Priors beta
for (i in 1:K) { beta[i] ~ dnorm(0, 0.0001)}
​
# Prior for delta parameter of GP distribution
delta ~ dunif(-1, 1)
​
C <- 10000
for (i in 1:N){
Zeros[i] ~ dpois(Zeros.mean[i])
Zeros.mean[i] <- -L[i] + C
l1[i] <- log(mu[i])
l2[i] <- (Y[i] - 1) * log(mu[i] + delta * Y[i])
l3[i] <- -mu[i] - delta * Y[i]
l4[i] <- -loggam(Y[i] + 1)
L[i] <- l1[i] + l2[i] + l3[i] + l4[i]
mu[i] <- (1 - delta)*exp(eta[i])
eta[i] <- inprod(beta[], X[i,])
}
​
# Discrepancy measures: mean, variance, Pearson residuals
for (i in 1:N){
ExpY[i] <- mu[i] / (1 - delta)
VarY[i] <- mu[i] / ((1 - delta)^3)
Pres[i] <- (Y[i] - ExpY[i]) / sqrt(VarY[i])
} }
",fill = TRUE)
​
sink()
​
inits <- function () {
list(beta = rnorm(ncol(X), 0, 0.1),
delta = 0)}
​
params <- c("beta", "delta")
​
GP1 <- jags(data = model.data,
inits = inits,
parameters = params,
model = "GP1reg.txt",
n.thin = 1,
n.chains = 3,
n.burnin = 4000,
n.iter = 5000)
​
print(GP1, intervals=c(0.025, 0.975), digits=3)
====================================================
Output on screen:
​
Inference for Bugs model at "GP1reg.txt", fit using jags,
3 chains, each with 5000 iterations (first 4000 discarded)
n.sims = 3000 iterations saved
mu.vect sd.vect 2.5% 97.5% Rhat n.eff
beta[1] 1.029 0.017 0.993 1.062 1.017 130
beta[2] 3.475 0.022 3.432 3.521 1.016 140
delta -0.287 0.028 -0.344 -0.233 1.002 1900
deviance 20005178.615 2.442 20005175.824 20005184.992 1.000 1
For each parameter, n.eff is a crude measure of effective sample size,
and Rhat is the potential scale reduction factor (at convergence, Rhat=1).
DIC info (using the rule, pD = var(deviance)/2)
pD = 3.0 and DIC = 20005181.6
DIC is an estimate of expected predictive error (lower deviance is better).