Parallel compute bayesian distribution regression

par_distreg uses parallel computation to compute bayesian distribution regression for a given vector of threshold values and data (with first column being the outcome variable)

par_distreg(thresh, data0, fn = distreg, no_cores = 1,
  type = "PSOCK", ...)

Arguments

thresh	vector of threshold values.
data0	the original data set with the dependent variable in the first column
fn	bayesian distribution regression function. the default is distreg provided in the package
no_cores	number of cores for parallel computation
type	type passed to makeCluster in the package parallel
...	any additional input parameters to pass to fn

Value

mat a G x M matrix of output (G is the length of thresh, M is the number of draws)

Examples

data0=faithful[,c(2,1)]; qnts<-quantile(data0[,1],c(0.05,0.25,0.5,0.75,0.95))
out<- par_distreg(qnts,data0,no_cores=1,iter = 102, burn = 2)
par(mfrow=c(3,2));invisible(apply(out,1,function(x)plot(density(x,30))));par(mfrow=c(1,1))