calcINT: calcINT

View source: R/calcINT.R

calcINTR Documentation

calcINT

Description

Marginalizing the likelihood through numerical integration.

Usage

calcINT(
  mlefit,
  lower,
  upper,
  reltol = 1,
  scale = 0,
  maxEval = 1000,
  verbose = FALSE
)

Arguments

mlefit

Fitted object using calcMLE

lower

Lower bounds of parameters. Must be in following order: mx1,..,mx_(nC-1),mu,sigma,beta,xi.

upper

Upper bounds of parameters. Must be in following order: mx1,..,mx_(nC-1),mu,sigma,beta,xi.

reltol

Required relative tolerance error of evaluations in integration routine.

scale

used to avoid underflow (should be the maximum likelihood value)

maxEval

Maximum number of evaluations in the adaptIntegrate function.

verbose

Whether printing limits to integrate over. Printing progress if maxEval>0

Value

ret A list(loglik,logdeviation,nEvals,scale) where loglik is Marginalized likelihood, logdeviation is the error-interval of loglik, nEvals is the total number of likelihood evaluations.

Author(s)

Oyvind Bleka

References

Hahn,T. (2005). CUBA - a library for multidimensional numerical integration. Computer Physics Communications, 168(2),78-95.


oyvble/euroformix documentation built on April 13, 2025, 3:18 a.m.