iprior: Characterizing Imprecise Prior

In imPois: Imprecise Inference for Poisson Sampling Models

Description

A set of prior distributions is characterized as an imprecise prior for inference. See ‘Details’.

Usage

iprior(ui, ci, pmat)

Arguments

ui	constraint matrix (k x p), see below.
ci	constrain vector of length k, see below.
pmat	matrix (k x p) containig coordinate information in d-dimensions.

Details

A convex set of hyperparameters is charcterized using a set of linear inequalities. grDevices::chull and geometry::convhulln functions are used to search for extreme points of this convex set.

Examples

## 2-dims (xi2=0, xi1, xi0)
lc0 <- list(lhs=rbind(diag(2), -diag(2)), rhs=c(0,0,-1,-1))
op <- iprior(ui=rbind(diag(2), -diag(2)), ci=c(0,0,-1,-1)) 
op <- iprior(ui=rbind(c(1,0),c(0,1),c(-1,-1)), ci=c(0,0,-5)) 
op <- iprior(ui=rbind(c(1,0),c(0,1),c(0,-1),c(1,1),c(-2,-1)), 
             ci=c(1,2,-8,5,-14)) # (3,8),(1,8), (1,4),(3,2)(6,2)

## 3-dimes (xi2, xi1, xi0)
op <- iprior(ui=rbind(c(1,0,0), c(-1,0,0), c(0,1,0), c(0,-1,0), c(0,0,1)), 
             ci=c(-0.5, -1, -2, -2, 0))
op <- iprior(ui=rbind(c(1,0), c(-1,0), c(0,1), c(0,-1)), 
             ci=c(0.5, -1, -2, -2))
lc0 <- cbind(rbind(c(1,0,0), c(-1,0,0), c(0,1,0), c(0,-1,0), c(0,0,1), 
             c(0,0,-1)), c(0.5, -1, -2, -2,0,-1))
iprior(pmat=lc0)
lc0 <- rbind(c(-2,1,0), c(2,1,0), c(-2,0.5,0), c(2,0.5,0))
lc0 <- rbind(c(1,2,0), c(1,-2,0), c(0.5,-2,0), c(0.5,2,0)) 
iprior(pmat=lc0)

imPois documentation built on May 30, 2017, 3:32 a.m.