cormat: Functions on correlation matrices

In vincenzocoia/CopulaModel: CopulaModel: Dependence Modeling with Copulas

Description

Functions on correlation matrices

Usage

corvec2mat(rvec) # convert correlation vector to matrix
cormat2vec(rmat) # extract vector from correlation matrix
corDis(Rmod,Robs,n=0,npar=0) # discrepancy between model-based and observed correlation matrices

Arguments

rvec	vector of length d*(d-1)/2 with order r[1,2],r[1,3],r[2,3],r[1,4], ..., r[d-1,d]
rmat	dxd correlation matrix
Rmod	model-based correlation matrix
Robs	observed correlation matrix
n	sample size for Robs
npar	parameter vector size leading to Rmod

Value

dxd correlation matrix for corvec2mat

vector of length d*(d-1)/2 for cormat2vec

log(det(Rmod))-log(det(Robs))+sum(diag(solve(Rmod,Robs)))-nrow(Robs) for corDis assuming a Gaussian dependence model

Examples

rvec=c(.3,.4,.5,.4,.6,.7)
Rmod=corvec2mat(rvec)
print(Rmod); print(chol(Rmod))
print(cormat2vec(Rmod))
robsvec=c(.32,.38,.53,.41,.61,.67)
Robs=corvec2mat(robsvec)
print(corDis(Rmod,Robs))
print(corDis(Rmod,Robs,n=400,npar=3))

vincenzocoia/CopulaModel documentation built on Oct. 27, 2021, 6:41 a.m.