R/joint.curvesCOP.R In wasquith/copBasic: General Bivariate Copula Theory and Many Utility Functions

```"joint.curvesCOP" <-
function(cop=NULL, para=NULL, type=c("and", "or"),
probs=c(0.5, 0.8, 0.90, 0.96, 0.98, 0.99, 0.995, 0.998),
zero2small=TRUE, small=1E-6, divisor=100, delu=0.001, ...) {

type <- match.arg(type)

"orfunc" <- function(v, u=NULL, LHS=NULL, cop=NULL, para=NULL, ...) {
RHS <- duCOP(u, v, cop = cop, para = para, ...)
#message("u=",u,",  v=",v, ",  LHS=",LHS, ",  RHS=",RHS, ", DELTA=", LHS-RHS)
return(LHS - RHS)
}

zz <- new.env()
for(t in probs) {
if(type == "and") {
# Concerning the tmp handling, there seems to be cases in which empirical copulas
# are being used with small sample sizes that can trigger a 'wrong sign in 'by' argument
# error on the following sequence, if so, let us try reversing the sequence and if that
# fails then bail out enterly.
delta <- diff(range(c(t+delu, 1-delu)))/divisor
tmp <- NULL
try(tmp <- seq(t+delu, 1-delu, by=delta), silent=FALSE)
if(is.null(tmp)) {
warning("trying to compensate for 'by' error by reversing")
try(tmp <- seq(t+delu, 1-delu, by=-delta), silent=FALSE)
}
u <- c(t, t+delu/100, t+delu/10, t+delu/5, t+delu/2,
tmp,
1-delu/2, 1-delu/5, 1-delu/10, 1-delu/100, 1)
u  <- sort(unique(u))
v  <- sapply(1:length(u), function(i) { COPinv(cop=cop, u[i], t, para=para) })
if(zero2small) v[v == 0] <-     small
if(zero2small) v[v == 1] <- 1 - small
uv <- data.frame(U=u, V=v)
assign(as.character(t), uv, envir=zz)
} else if(type == "or") {
delta <- diff(range(c(0, t)))/divisor
u <- c(0, delu/100, delu/10, delu/5, delu/2,
seq(delu,t, by=delta),
t-delu/2, t-delu/5, t-delu/10, t-delu/100, t)
u <- sort(unique(u))
v <- vector(mode="numeric", length(u))
for(i in 1:length(u)) {
lo <- duCOP(0,0, cop=cop, para=para, ...)
if(is.na(lo) | ! is.finite(lo)) lo <- .Machine\$double.eps
my.rt <- NULL
try(my.rt <- uniroot(orfunc, interval=c(lo,t), u=u[i], LHS=t,
cop=cop, para=para, ...), silent=TRUE)
if(is.null(my.rt)) {
v[i] <- NA
} else if(length(my.rt\$root) != 0) {
v[i] <- my.rt\$root
} else {
v[i] <- NA
}
}
if(zero2small) v[v == 0] <-     small
if(zero2small) v[v == 1] <- 1 - small
uv <- data.frame(U=u, V=v)
assign(as.character(t), uv, envir=zz)
} else {
stop("should not be here in logic")
}
}
zzz <- as.list(zz); rm(zz)
return(zzz)
}
```
wasquith/copBasic documentation built on Nov. 20, 2019, 5:12 a.m.