[Rd] inappropriate definition of kurtosis in e1071 package (PR#7309)

From: <rafaminos_at_yahoo.fr>
Date: Sun 24 Oct 2004 - 02:14:06 EST


Full_Name: Raphal Césari
Version: 1.9.1
OS: Windows XP family edition
Submission from: (NULL) (213.100.40.112)

Hello,

I am a student at the Departments of Economics at Stockholm University.I am a beginner in Econometrics and was very confused when using the built in command "kurtosis()" of the e1071 package.When I computed the test,I was thinking of getting the kurtosis of my distribution which has to be greater or equal to zero,and I got a value under zero.The formula of kurtosis is :

     If N = length(x), then the kurtosis of x is defined as

            N^(-1) sd(x)^(-4) sum_i (x_i - mean(x))^4

and the function computes what i think is called the excess kurtosis which is just the same formula minus 3.This is why it is possible to get a computed value under zero!Now it is trivial to me,because I know the function but t would be good in the details of the function to explain that the computed value is the excess kurtosis which i think is more appropriate.

Best regards,

Raphaël Césari



R-devel@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel Received on Sun Oct 24 02:19:31 2004

This archive was generated by hypermail 2.1.8 : Fri 18 Mar 2005 - 09:00:56 EST