predict.qde: Posterior predictive summary for quantile-based density...

In qrjoint: Joint Estimation in Linear Quantile Regression

Posterior predictive summary for quantile-based density estimation

Description

Extract posterior predictive density estimate for qde

Usage

 
## S3 method for class 'qde'
predict(object, burn.perc = 0.5, nmc = 200, yRange = range(object$y), yLength = 401, ...)

Arguments

object	a fitted model of the class 'qde'.
burn.perc	a positive fraction indicating what fraction of the saved draws are to be discarded as burn-in
nmc	integer giving the number of samples, post burn-in, to be used in Monte Carlo averaging
yRange	range of values over which posterior predictive density is to be evaluated
yLength	number of grid points spanning yRange for posterior predictive density evaluation
...	currently no additional arguments are allowed

Value

Returns a list with three items:

y	vector giving the grid over which the posterior predictive density is evaluated.
fsamp	a matrix with yLength many rows and nmc many columns. Each column corresponds to a draw of the response density from the posterior predictive.
fest	summary of the posterior predictive density given by point-wise median, 2.5th and 97.5th percentiles.

See Also

qde and summary.qde.

Examples

 
# Plasma data analysis

data(plasma)
Y <- plasma$BetaPlasma
Y <- Y + 0.1 * rnorm(length(Y)) ## remove atomicity

# model fitting with 50 posterior samples from 100 iterations (thin = 2)
fit.qde <- qde(Y, 50, 2)
pred <- predict(fit.qde)
hist(Y, freq = FALSE, col = "gray", border = "white", ylim = c(0, max(pred$fest)))
matplot(pred$y, pred$fest, type="l", col=1, lty=c(2,1,2), ylab="Density", xlab="y")

qrjoint documentation built on April 6, 2023, 1:07 a.m.