log_normalize: Normalize in log space

In pauwels2014: Bayesian Experimental Design for Systems Biology.

Description

Binding to a C implementation of addition and normalization in log space.

Usage

log_normalize(vals, maxi = max(vals), ID = which.max(vals)[1])

Arguments

vals	Log-values to be normalize
maxi	The maximum value of vals
ID	One entry of vals which has the highest value.

Value

A list with three slots. First one is a numerical vector with values normalized so that they sum to 1. The second is the same normalized vector in log space. The last one is the log of the normalization constant. See example.

Author(s)

Edouard Pauwels

Examples

log_x <- rnorm(10)

temp <- log_normalize(log_x)

exp(temp[[2]]) - temp[[1]]

sum(exp(temp[[2]]))

log_x - temp[[3]] - temp[[2]]

pauwels2014 documentation built on May 1, 2019, 6:29 p.m.