log_f_logden: compute the log of density of inactive states' probability...

In zhenkewu/rewind: Reconstructing Etiology with Binary Decomposition

Description

This function is used in sampling a model with infinite number of columns in H

Usage

log_f_logden(log_p0, alpha, t)

Arguments

log_p0	log probability (the argument of this function)
alpha	hyperparameter for Indian Buffet Process (Infinite version)
t	the number of rows in the IBP

Value

a value corresponding to the log density f(log_p0)

Examples

library(ars)
n_grid  <- 10000
p0_grid <- seq(log(0.001),0,len=n_grid)
y_den  <- log_f_logden(p0_grid,5,8)

y <- ars::ars(n_grid,log_f_logden,log_f_logprima,
c(-6,-4,-2,-1),m=4,
ub=TRUE,xub=0,alpha=5,t=8)
hist(y,breaks="Scott",main='Adaptive Rjection Sampling',freq=FALSE)
rug(y)
points(density(y),type="l",col="blue")
points(p0_grid,exp(y_den-rewind:::logsumexp(y_den)-log(diff(p0_grid)[2])),
      type="l",col="red") # <-- true density; log_f_logden is only
                         #  correct upto a proportionality constant.

legend("topleft",c("Sample Density","True Density"),lty=c(1,1),col=c("blue","red"),
      bty="n")

zhenkewu/rewind documentation built on Sept. 9, 2020, 3:40 p.m.