| dzigp {ZIGP} | R Documentation |
'dzigp' calculates the probability mass function of the ZIGP distribution.
dzigp(x, mu, phi, omega)
x |
vector of discrete points |
mu |
mean |
phi |
dispersion parameter |
omega |
zero inflation parameter |
Calculates a vector of the same length as of x of pmf-values.
x <- 1:10
dzigp(x, 2, 1.5, 0.2)
#[1] 0.201467310 0.144357635 0.093859287 0.058560538 0.035846689 0.021749120
#[7] 0.013146607 0.007938715 0.004796227 0.002901477
## The function is currently defined as
function(x, mu = stop("no mu arg"), phi = stop("no phi arg"),
omega = stop("no omega arg")){
# check if parameters are valid
if(omega < 0) {return("omega has to be in [0,1]!")}
if(omega > 1) {return("omega has to be in [0,1]!")}
upper <- max(x)
p <- double(upper+1)
#P(X=0)
p[1] <- omega + (1-omega) * exp(-mu/phi)
if (upper > 0) {
rekursive <- FALSE
for (i in 1:upper) {
#P(X=x)
if (rekursive==FALSE) {
p[i+1] <- (1-omega)*mu*(mu+(phi-1)*i)^(i-1)/exp(lgamma(i+1))*
phi^(-i)*exp(-1/phi*(mu+(phi-1)*i))}
if (p[i+1]==Inf) {
rekursive <- TRUE
log.p.alt <- log( (1-omega)*mu*(mu+(phi-1)*(i-1))^(i-2)/
exp(lgamma(i-1+1))*
phi^(-(i-1))*exp(-1/phi*(mu+(phi-1)*(i-1))))
}
if (rekursive==TRUE) {
log.p <- log( (mu+(i-1)*(phi-1))/(phi*i)*
(1+(phi-1)/(mu+(i-1)*(phi-1)))^(i-1)*
exp(1/phi-1) ) + log.p.alt
log.p.alt <- log.p
p[i+1] <- exp(log.p)
}
}
}
p2 <- double(length(x))
for (i in 1:length(x)) {
p2[i] <- p[x[i]+1]
}
return(p2)
}