jac.logit: Jacobian for logit transform

In bisque: Approximate Bayesian Inference via Sparse Grid Quadrature Evaluation (BISQuE) for Hierarchical Models

Description

Let X=logit(Y) be a transformation of a random variable Y that lies in the closed interval (L,U). This function computes the jacobian J(x) when using the density of Y to evaluate the density of X via

f(x) = f_y(logit^{-1}(x) * (U-L) + L) J(x)

where

J(x) = (U-L) d/dx logit^{-1}(x).

Usage

jac.logit(x, log = TRUE, range = c(0, 1))

Arguments

x	value at which to evaluate J(x)
log	TRUE to return log(J(x))
range	vector specifying min and max range of the closed interval for the logit. While the logit is defined for real numbers in the unit interval, we extend it to real numbers in arbitrary closed intervals (L,U).

Examples

jac.logit(1)

bisque documentation built on March 26, 2020, 7:27 p.m.