scdeco.cop: Copula dynamic correlation fitting function

In scDECO: Estimating Dynamic Correlation

Copula dynamic correlation fitting function

Description

Copula dynamic correlation fitting function

Usage

scdeco.cop(
  y,
  x,
  marginals,
  w = NULL,
  n.mcmc = 10000,
  burn = 1000,
  thin = 1,
  offset1 = NULL,
  offset2 = NULL
)

Arguments

y	2-column matrix of observations
x	covariates
marginals	length-2 vector with strings of the two marginals
w	(optional)
n.mcmc	number of mcmc iterations to run
burn	how many of the mcmc iterations to burn
thin	how much to thin the mcmc iterations
offset1	(optional) offset for link(mu1)
offset2	(optional) offset for link(mu2)

Value

matrix with mcmc samples as rows and columns corresponding to the different parameters

Examples

n <- 1000
x.use = rnorm(n)
w.use = runif(n,-1,1)
eta1.use = c(-2.2, 0.7)
eta2.use = c(-2, 0.8)
beta1.use = c(1,0.5)
beta2.use = c(1,1)
alpha1.use = 7
alpha2.use = 3
tau.use = c(-0.2, .3)

marginals.use <- c("ZINB", "ZIGA")

y.use <- scdeco.sim.cop(marginals=marginals.use, x=x.use,
                    eta1.true=eta1.use, eta2.true=eta2.use,
                    beta1.true=beta1.use, beta2.true=beta2.use,
                    alpha1.true=alpha1.use, alpha2.true=alpha2.use,
                    tau.true=tau.use, w=w.use)
mcmc.out <- scdeco.cop(y=y.use, x=x.use, marginals=marginals.use, w=w.use,
                      n.mcmc=10, burn=0, thin=1) # n.mcmc=1000, burn=100, thin=5)

lowerupper <- t(apply(mcmc.out, 2, quantile, c(0.025, 0.5, 0.975)))
estmat <- cbind(lowerupper[,1],
                c(eta1.use, eta2.use, beta1.use, beta2.use, alpha1.use, alpha2.use, tau.use),
                lowerupper[,c(2,3)])
colnames(estmat) <- c("lower", "trueval", "estval", "upper")
estmat

scDECO documentation built on April 12, 2025, 2:11 a.m.