Learn R Programming

FactorCopulaModel (version 0.1.1)

Factor Copula Models

Description

Inference methods for factor copula models for continuous data in Krupskii and Joe (2013) , Krupskii and Joe (2015) , Fan and Joe (2024) , one factor truncated vine models in Joe (2018) , and Gaussian oblique factor models. Functions for computing tail-weighted dependence measures in Lee, Joe and Krupskii (2018) and estimating tail dependence parameter.

Copy Link

Version

Install

install.packages('FactorCopulaModel')

Monthly Downloads

147

Version

0.1.1

License

GPL-3

Maintainer

Pavel Krupskii

Last Published

November 6th, 2025

Functions in FactorCopulaModel (0.1.1)

Kendall's tau for bivariate normal

Rank-based normal scores transform

Compute correlation matrix according to 1-factor + 1-truncated vine (residual dependence) model

R interface for Gauss-Legendre quadrature

nestfactorcop_nllk

negative log-likelihoods of nested factor structured factor copula and derivatives computed in f90 for input to posDefHessMinb

simulate from bi-factor copula model

negative log-likelihood for the p-factor Gaussian/t model

Gaussian oblique factor structure correlation matrix

correlation matrix for 1-factor plus 1-truncated vine (for residual dependence)

oblique_pp_par2load

oblique factor correlation structure for d variables and m groups include determinant and inverse

onefactorEstWithProxy

Parameter estimation for 1-factor copula with estimated latent variables using VineCopula::BiCopSeelct

C_[2|1]^[-1](p|u) for bivariate Student t copula

log-likelihood Gaussian p-factor structure correlation matrix

Gaussian p-factor structure correlation matrix

C_[2|1]^[-1](p|u) for bivariate Frank copula

frank_beta2cpar

Frank: Blomqvist's beta to copula parameter

log-likelihood Gaussian bi-factor structure correlation matrix

Plot zeta(alpha) against alpha

bifactorcop_nllk

negative log-likelihood of bi-factor structured factor copula and derivatives computed in f90 for input to posDefHessMinb

simulate from 1-factor copula model with different linking copula families

Precipitation by rainstorm at 28 stations

negative log-likelihood for the bi-factor Gaussian/t model

gumbel_rhoS2cpar

Gumbel: Spearman rho to copula parameter

gumbel_beta2cpar

Gumbel: Blomqvist's beta to copula parameter

Integrand for 1-factor copula with 1-parameter bivariate linking copula families; or for m-parameter bivariate linking copulas

lower and upper bounds for copula parameters (1-parameter, 2-parameter families)

max likelihood (min negative log-likelihood) for 1-factor copula model

min negative log-likelihood for 1-factor copula with nlm()

oblique_grad_fa

Gaussian oblique factor structure correlation matrix

Empirical version of zeta(alpha) tail-weighted dependence measure

MLE in a MVt model with a p-factor correlation structure

Spearman's rho for bivariate copula with parameter cpar

oblique_grad_nllk

log-likelihood Gaussian oblique factor structure correlation matrix

Random multivariate normal (standard N(0,1) margins)

Semi-correlations for two variables

Semi-correlation table for a multivariate data set

frank_rhoS2cpar

Frank: Spearman rho to copula parameter

Upper Tail-weighted dependence measure zeta(C,alpha)

oblique_par2load

oblique factor correlation structure for d variables and m groups

Tail dependence parameter estimation

log-likelihood Gaussian oblique factor structure correlation matrix

Minimization with modified Newton-Raphson iterations, Hessian is modified to be positive definite at each step. Algorithm and code produced by Pavel Krupskii (2013) see PhD thesis Krupskii (2014), UBC and Section 6.2 of # Joe (2014) Dependence Models with Copulas. Chapman&Hall/CRC

Version with ifixed as argument

Rank-based uniform scores transform

MLE for multivariate normal/t with a bi-factor or nested factor correlation structure

min negative log-likelihood for 1-factor copula model (some parameters can be fixed)

onefactorcop_nllk

negative log-likelihood of 1-factor copula for input to posDefHessMin and posDefHessMinb

Partial correlation representation to loadings for p-factor

Random multivariate t (standard t(nu) margins)

Simulate data from nested copula or Gaussian model

BB1, given 0<tau<1, find theta and delta with lower tail dependence equal upper tail dependence

BB1 copula parameter (theta,delta) to tail dependence parameters

Proxies for bi-factor copula model based on Gaussian bi-factor score

Gaussian bi-factor structure correlation matrix

GARCH-filtered log returns for Dow Jones stocks 2014-2016

bifactorEstWithProxy

Sequential parameter estimation for bi-factor copula with estimated latent variables using VineCopula::BiCopSelect

Bi-factor partial correlations to correlation matrix

bifactor2cor_v2

Bi-factor partial correlations to correlation matrix version 2, using the inverse and determinant of a smaller matrix

compute correlation matrix from 2-truncated R-vine

BB1 tail dependence parameters to copula parameter (theta,delta)

Convert from correlations in vector form to a correlation matrix

Discrepancy of model-based and observed correlation matrices based on Gaussian log-likelihood

log returns and GARCH-filtered log returns for some Euro markets 2007

latentUpdate1factor

Compute new proxies for 1-factor copula based on the mean of observations

Check if a square symmetric matrix is positive definite

factor1trvine_nllk

negative log-likelihood with gradient and Hessian computed in f90 for copula from 1-factor/1-truncated vine (tree for residual dependence conditional on a latent variable); models included are BB1 for latent with Frank or Gaussian(bvncop) for truncated vine residual dependence

latentUpdateBifactor

Conditional expectation proxies for bi-factor copula models with linking copulas in different copula families

latentUpdate1factor1

Compute new proxies for 1-factor copula based on the mean of observations

Semi-correlation for bivariate normal/Gaussian distribution