Set_Grid: Calculate Log-densities for Grid Components

Description Usage Arguments Value Examples

View source: R/EnvSet_Grid.R

Description

Computes un-normalized log of the density associated with each component of a Grid used to sample using the likelihood-subgradient density approach

Usage

1
Set_Grid(GridIndex, cbars, Lint)

Arguments

GridIndex

A matrix containing information for each component in the grid related to whether the components is to the left, in the center, or to the right of the density. Each row corresponds to a component in the grid, while the columns correspond to the transformed (standardized) variables.

cbars

A matrix containing the subgradient for the (adjusted) negative log-likelihood function for each the component in the grid.

Lint

A matrix storing information on where the upper and lower bounds are, depending on whether the sampling is from the left, the center, or the right.

Value

Refer to Nygren and Nygren (2006) for details. The first set of items refers to Example 2 in section 3.1. All except the last item in this list of returned items has a number of rows equaling the number of components of the grid and a number of columns equaling the number of coefficients in the model. All quantities refer to the respective coefficient for each of the components of the grid.

Down

The lower bounds for the interval to be evaluated. Either negative infinity or a real number.

Up

The upper bounds for the interval to be evaluated. Either positive infinity or a real number.

lglt

The log of the density between negative infinity and the upper bound

lgrt

The log of the density between the lower bound and infinity

lgct

The log of the density between the lower and upper bounds

logU

The one of the 3 above that is relevant for the component of the grid

logP

A two column matrix, the first of which holds sum of logU across the components. The second column is 0 and is later populated by the Set_logP function

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
## ----dobson-------------------------------------------------------------------
## Dobson (1990) Page 93: Randomized Controlled Trial :
counts <- c(18,17,15,20,10,20,25,13,12)
outcome <- gl(3,1,9)
treatment <- gl(3,3)

## Prior mean vector 
mu<-matrix(0,5)           
mu[1,1]=log(mean(counts)) 
## Prior standard deviation and Variance
mysd<-1           
V=((mysd)^2)*diag(5)  
## Call to glmb
glmb.D93<-glmb(n=1000,counts ~ outcome + treatment,
               family = poisson(),pfamily=dNormal(mu=mu,Sigma=V))
## ----glmb extractAIC-------------------------------------------------------------
extractAIC(glmb.D93)

knygren/glmbayes documentation built on Sept. 4, 2020, 4:39 p.m.