waic: Watanabe Information Criterion
In qrjoint: Joint Estimation in Linear Quantile Regression

Description Usage Arguments Value References See Also Examples

View source: R/qrjoint.R

Calculates two versions of the Watanabe information criteria from MCMC draws.

1 2	waic(logliks, print = TRUE)

`logliks`	a matrix of observation level log-likelihood values, the columns are MCMC iterations and the rows are observations in the data
`print`	logical whether to print the results

Returns the two version of the WAIC

Gelman, A., Hwang, J., and Vehtari, A. (2014). Understanding predictive information criterion for Bayesian models. Stat Comput, 24, 997-1016.

summary.qrjoint

 
# Plasma data analysis

# recoding variables
data(plasma)
plasma$Sex <- as.factor(plasma$Sex)
plasma$SmokStat <- as.factor(plasma$SmokStat)
plasma$VitUse <- 3 - plasma$VitUse
plasma$VitUse <- as.factor(plasma$VitUse)

# creating predictors and response (beta carotene concentration in the plasma)
X <- model.matrix(BetaPlasma ~ Age + Sex + SmokStat + Quetelet + VitUse + Calories + 
        Fat + Fiber + Alcohol + Cholesterol + BetaDiet, data = plasma)[,-1]
Y <- plasma$BetaPlasma

# model fitting with 50 posterior samples from 100 iterations (thin = 2)
fit.qrj <- qrjoint(X, Y, 50, 2)
sm <- summary(fit.qrj, plot = FALSE)

# the call to summary already shows the waic for the fitted model, it also returns 
# the observation level log-likelihood vales. To calculate waic from last 20 draws 
# we can use:

ll <- sm$ll
waic(ll[,31:50])