rstandard.fitlgc {lgc}R Documentation

Simulation based residuals for goodness of fit.

Description

Generate an approximate posterior sample standardized with the prior.

Usage

  rstandard.fitlgc(model, fixbeta = TRUE, seed = 123)

Arguments

model

A fitted model object.

fixbeta

Logical. Should the fixed effects be deterministic in the sample?

seed

Simulation seed.

Details

A goodness of fit test for the log Gaussian Cox model can be performed using the following simple principle: Draw one sample η^* from the posterior distribution η|x. Standardize with the prior u=Σ^{-1/2}η^*. If the prior is correctly specified then u is a vector of independent standard normal variates - which can be further investigated by a Kolmogorov-Smirnov test or a Chi-square test. The simulation seed is deliberately fixed to stress the assumption of one sample.

Value

A list with a vector of standardized residuals for each model term.


[Package lgc version 1.4 Index]