makedeptable: Make table of dependence measures for a 1-parameter bivariate...

In YafeiXu/CopulaModel: CopulaModel: Dependence Modeling with Copulas

Description

Make table of dependence measures for a 1-parameter bivariate copula family

Usage

makedeptable(bvec,bfn,dcop,pcop,pcond12,pcond21,LBcpar=0,UBcpar=Inf,
   itaildep=F,lmfn,zero=0,zbd=6,iprint=F)

Arguments

bvec	vector of Blomqvist beta values in increasing order, first is 0, last is 1
bfn	function to get copula parameter cpar given Blomqvist beta
dcop	function for copula density c
pcop	function for copula cdf C
pcond12	function for copula conditional C_{1|2}(u|v)
pcond21	function for copula conditional C_{2|1}(v|u)
LBcpar	parameter for independence
UBcpar	parameter for comonotonicity
itaildep	T to compute tail dependence parameter lambda=lm
lmfn	function to compute lambda given copula parameter
zero	0 or something like 1.e-6, used with pcond12, pcond21
zbd	integration bound with respect to N(0,1) margins for rhoN
iprint	print flag for intermediate results, default=F

Value

table with column headings of "cpar","beta","tau","rhoS","rhoN","lambda"

See Also

deptab depmeas2cpar

Examples

bvec=seq(0,.9,.02)  # more grid points
bvec=c(bvec,.95)
bvec=c(bvec,1)
np=length(bvec)
bvec=bvec[-(np-1)]  # 0.95 can be a problem for some copula families
np=length(bvec)
frk.deptab=makedeptable(bvec,bfn=frk.b2cpar,pcop=pfrk,
  pcond12=pcondfrk,pcond21=pcondfrk,LBcpar=0,
  UBcpar=Inf,itaildep=FALSE,zero=0,zbd=7,iprint=TRUE)
tauinteg=frk.deptab[,3]
for(i in 2:(np-1))
{ frk.deptab[i,3]=frk.cpar2tau(frk.deptab[i,1]) }
print(frk.deptab)
cat("accuracy of 2-dimensional numerical integration\n")
print(abs(tauinteg-frk.deptab[,3]))

YafeiXu/CopulaModel documentation built on May 9, 2019, 11:07 p.m.