[R] Additional args to fun in integrate() not found?
Sergey Goriatchev
sergeyg at gmail.com
Tue Feb 27 10:49:22 CET 2007
Hello, fellow Rdicts,
I have the code for the program below. I need to integrate a function
of "x" and "p". I use integrate to integrate over "x" and pass "p" as
an additional argument. "p" is specified and given default value in
the argument list. Still, integrate() cannot read "p", unless I
explicitly insert a numeric value in the integrate() argument list.
And when I do that, I get the right result, but still some warnings.
Please, help me with these problems:
1) why is "p" not recognized?
2) what are these warning messages?
PROGRAM CODE:
---------------------------
#THIS LIBRARY IS NEEDED FOR THE INCOMPLETE GAMMA FUNCTION
library(zipfR)
#------------------------------------------------------------------------
gedCDF = function(yvec, p=2, mu=0, scale=1, numint=0)
{
#-------------------------------------------------------------------------
#Setting k to sqrt(2) and the GED with p=2 coincides with standard normal.
#Set k=1 and GED with p=1 coincides with Laplace.
k<-sqrt(2)
#k<-1
scale<-scale*k
zvec<-(yvec-mu)/scale
cdf<-matrix(0, length(zvec),1)
for(i in 1:length(zvec))
{
z<-zvec[i]
if(numint==0)
{
if(z<=0)
{
t<-0.5*(1-1/gamma(1/p)*Igamma((1/p),(-z)^p,lower=TRUE))
}
else
{
t<-1-(0.5*(1-1/gamma(1/p)*Igamma((1/p),(z)^p,lower=TRUE )))
}
}
else
{
t<-integrate(geddenstandard, -35, z, subdivisions=1000,
rel.tol=100*.Machine$double.eps, abs.tol=rel.tol,
stop.on.error=TRUE, keep.xy=FALSE, aux=NULL,p)
}
cdf[i]<-t
}
cdf
}
#-------------------------------------------------------------------------
geddenstandard = function(z,p)
{
f<-p/(2*gamma(1/p))*exp(-abs(z)^p)
}
-----------------------------------------------------------------------------------------------------------------------
If I run with this definition I get the following error message and abort:
> gedCDF(c(1,2,3,4,5), numint=1)
Error in eval(expr, envir, enclos) : ..1 used in an incorrect context,
no ... to look in
If I replace "p" in integrate() with 2, I get correct answers, but
still some warning messages:
> gedCDF(c(1,2,3,4,5), numint=1)
[[1]]
[1] 0.8413447
[[2]]
[1] 0.9772499
[[3]]
[1] 0.9986501
[[4]]
[1] 0.9999683
[[5]]
[1] 0.9999997
Warning messages:
1: number of items to replace is not a multiple of replacement length
2: number of items to replace is not a multiple of replacement length
3: number of items to replace is not a multiple of replacement length
4: number of items to replace is not a multiple of replacement length
5: number of items to replace is not a multiple of replacement length
-------------------------------
Do I get these warnings because I define cdf as a matrix and the
output-cdf is a list?
Please, help me with these!
Email to my gmail account, please: sergeyg at gmail.com
THanks in advance
Sergey
More information about the R-help
mailing list