FactorCopulaModel (version 0.1.1)

posDefHessMinb: Version with ifixed as argument

Description

modified Newton-Raphson minimization with positive Hessian

Usage

posDefHessMinb(param,objfn,ifixed,dstruct,LB,UB,mxiter=30,eps=1.e-6,
  bdd=5,iprint=FALSE)

Value

list withfnval = function value at minimum; parmin = param for minimum; invh = inverse Hessian; iconv = 1 if converged, -1 for a boundary point, 0 otherwise; iter = number of iterations.

Arguments

param

starting point for minimization

objfn

function to be minimized with gradient and Hessian

ifixed

vector of length(param) of TRUE/FALSE, such that ifixed[i]=TRUE iff param[i] is fixed at the given value

dstruct

list with data set and other variables used by objfn

LB

lower bound vector

UB

upper bound vector

mxiter

max number of iterations

eps

tolerance for Newton-Raphson iterations

bdd

bound on difference of 2 consecutive iterations (useful is starting point is far from solution and func is far from convex)

iprint

control on amount of printing, FALSE for no printing of iterations and TRUE for printing x^(k) on each iteration.

Examples

# See examples in onefactorcop_nllk(), bifactorcop_nllk(), nestfactorcop_nllk()