Skip to content

package(moments) issue

5 messages · Bill Dunlap, Sania Wadud

#
Hi all,

While running the anscombe.test in R, I'm getting an error of *Error in if
(pval > 1) pval <- 2 - pval : missing value where TRUE/FALSE needed* for a
few time series columns whereas for most of the series the function is
working fine. I have checked for those specific columns for missing values.
However, there is no NA/NAN value in the dataset.

I have also run kurtosis for those time series and I can get results. I
have noticed the same issue was reported in the following post for python
(scipy.stat) and the report mentions that there is a bug in moments package
in R. I was wondering whether the issue has been resolved.

https://github.com/scipy/scipy/issues/1950

Could anyone please help me with the error?

I look forward to hearing from you.

Regards,
Sania
#
moments::anscombe.test(x) does give errors when x has too few values or if
all the values in x are the same
Error in if (pval > 1) pval <- 2 - pval :
  missing value where TRUE/FALSE needed
Error in if (pval > 1) pval <- 2 - pval :
  missing value where TRUE/FALSE needed

You can use tryCatch() to return some special value where these errors
occur.  E.g., use

tryCatch(anscombe.test(x), error=function(e) list(p.value=NA))

instead of ascombe.test (this assumes you will only look at p.value later -
add other entries as needed.).

-Bill
On Thu, Oct 15, 2020 at 12:09 PM Sania Wadud <sania.wadud at gmail.com> wrote:

            

  
  
#
Hi Bill,

Thanks for prompt reply and letting me know a way around it.

I have more than 1200 observations and not all the values are the same.
However, my data points are quite similar, for example,
0.079275, 0.078867, 0.070716  in millions and etc. I have run the data
without converting it to millions and I still get the same error
message. As I have kurtosis value, it should be fine for the time being.
However, in case if you have any other explanation behind the
error, I will highly appreciate it.

Best wishes,
Sania
On Thu, 15 Oct 2020 at 20:43, Bill Dunlap <williamwdunlap at gmail.com> wrote:

            

  
  
#
Another bad case is
    > moments::anscombe.test(rep(c(1,1.1),length=35))
    Error in if (pval > 1) pval <- 2 - pval :
      missing value where TRUE/FALSE needed

I haven't checked the formulas carefully, but I suspect the problem is from
taking the cube root of a negative number in
    z <- (1 - 2/(9 * a) - ((1 - 2/a)/(1 + xx * sqrt(2/(a -
4))))^(1/3))/sqrt(2/(9 * a))
In R, the cube root (or any non-integral power) of a negative real number
is NaN and  ((1 - 2/a)/(1 + xx * sqrt(2/(a - 4))))
is c. -928 in this example.  I think the intent was for the cube root to be
a negative real, c. -9.76, in this case.

If that is the intent, then a fix is to define a cube_root function for
reals
   cube_root <- function (x) sign(x) * abs(x)^(1/3)
and change the z<-... line to use cube_root() instead of ^(1/3).
    z <- (1 - 2/(9 * a) - cube_root(((1 - 2/a)/(1 + xx * sqrt(2/(a -
4))))))/sqrt(2/(9 * a))

-Bill
On Thu, Oct 15, 2020 at 3:15 PM Sania Wadud <sania.wadud at gmail.com> wrote:

            

  
  
#
Hi Bill,

Thanks for rectifying the issue. It seems to be a cube root of a negative
real issue. After updating the function it's working fine
and showing the correct results.

I highly appreciate your help.

Best wishes,
Sania
On Fri, 16 Oct 2020 at 00:36, Bill Dunlap <williamwdunlap at gmail.com> wrote: