factorcopsim: Simulation for factor copulas

In YafeiXu/CopulaModel: CopulaModel: Dependence Modeling with Copulas

Description

Simulation for 1-factor copula and 2-factor copula, all bivariate linking copulas in same parametric family for each factor

Usage

sim1fact(n,parobj1,qcond1,copname1,ivect=F)
sim2fact(n,parobj1,parobj2,qcond1,qcond2,copname1="a1",copname2="a2",ivect=F)

Arguments

n	sample size
parobj1	parameter vector of dimension d or parameter matrix with d rows, where d is dimension of factor copula; parobj is dx2 for something like BB1 copula
parobj2	parameter vector of dimension d or parameter matrix with d rows, for linking copulas to factor 2
qcond1	name of conditional copula function C_{U|V}^{-1}(u|v), choices include qfrk, qgum, qgumr, qbb1, qt with fixed nu1.
qcond2	name of conditional copula function C_{U|V}^{-1}(u|v) for second factor, choices include qfrk, qgum, qgumr, qt with fixed nu2.
copname1	copula name: the function checks on "frank", "mtcj", "mtcjr", "fgm" for which qcond has closed form.
copname2	copula name for factor 2
ivect	flag that is T if qcond1 and qcond2 have vectorized forms

Value

data matrix of dimension nxd

References

Krupskii P and Joe H (2013). Factor copula models for multivariate data. Journal of Multivariate Analysis, 120, 85-101.

See Also

factorcopmle rfactcop

Examples

 
bevec=c(.8,.7,.6,.5,.5)
cpar.frk=frk.b2cpar(bevec)
lmbb1=matrix(c(.3,.4,.5,.3,.5, .6,.6,.6,.7,.7),5,2)
cpar.bb1=lmbb1
for(i in 1:nrow(lmbb1))
{ cpar.bb1[i,]=bb1.lm2cpar(lmbb1[i,]) }
n=300
set.seed(123)
frkdat=sim1fact(n,cpar.frk,qcondfrk,"frank")
print(cor(frkdat))
set.seed(123)
bb1dat=sim1fact(n,cpar.bb1,qcondbb1,"bb1",ivect=FALSE)
print(cor(bb1dat))
set.seed(123)
bb1dat=sim1fact(n,cpar.bb1,qcondbb1,"bb1",ivect=TRUE)
print(cor(bb1dat))
# pairs(bb1dat)
bevec2=c(.0001,.6,.6,.6,.7)
cpar.frk2=frk.b2cpar(bevec2)
frk2dat=sim2fact(n,cpar.frk,cpar.frk2,qcond1=qcondfrk,qcond2=qcondfrk,"frank","frank")
print(cor(frk2dat))

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