sas >> Kolmogorov or Shapiro?

by sasuser.joey.m » Thu, 22 Oct 2009 11:11:48 GMT

Hi guys:

I have read that Shapiro-Wilk is recommended when N is less than 2000, and
Kolmogorov is better when N is larger.

Do you have an advice? a good recommendation for a reading?

Have a wonderful day (actually night, in my side of the world).

Cheers.

Joey.


sas >> Kolmogorov or Shapiro?

by sasuser.joey.m » Thu, 22 Oct 2009 11:15:39 GMT


Sorry, (it may be obvious for many of you) but I did not mention that I want
to test for normality.

J.







sas >> Kolmogorov or Shapiro?

by goladin » Thu, 22 Oct 2009 11:19:31 GMT

Hi,

How abt anderson-darling test?

------Original Message------
From: joey m
Sender: SAS(r) Discussion
To: XXXX@XXXXX.COM
ReplyTo: joey m
Subject: Re: Kolmogorov or Shapiro?
Sent: Oct 22, 2009 11:15 AM

Sorry, (it may be obvious for many of you) but I did not mention that I want
to test for normality.

J.






Sent from my BlackBerry Wireless Handheld

Certified Advanced Programmer for SAS V9
Certified Basic Programmer for SAS V9


Similar Threads

1. Shapiro wilks test

Hi All,


Does anybody have a detail document on shapiro wilks test or documentation
for the same.I want to get an in deep idea about its concept.

Thanks in advance

Ankur Arora

2. Shapiro and Wilks test of normailty: bilateral or unilateral

3. Kolmogorov - Smirnov Test for uniform distribution

Dear SAS users,

I think this is a simple question, but I somewhat get stucked, don't
know how to do it.

I try to use SAS Npar1way procedure to test if my variable with 100
observations is a uniform distribution(0,1).

What I did is to simulate a random variable with 100 observations from
uniform distriubtion, then compare my variable to simulated variable
using proc npar1way.

data mydata;
   set mydata;
   dist=1;
   rename probt=pvalue;
run;

data uniform;
do i=1 to 100;
pvalue=RANUNI(8765);
dist=2;
output;
end;
run;

data compare;
set mydata uniform;
run;

proc npar1way data=compare EDF D;
class dist;
var pvalue;
run;

But since my random variable is simulated, different seed will produce
different p-value. Does anyone of you know how to test my data against
a uniform distribution without doing simulation?

Thanks in advance.

Qi

4. Kolmogorov-Smirnov Statistic--What is it?

5. Kolmogorov-Smirnov D

One of our clients sent over a (logistic) predictive model validation chart,
in which a D was calculated. It is max {|edf(1) - edf(all)|}, with edf(1)
being the cumulative percent of responders and edf(all) the cumulative
percent of total mail volume (responders + non-responders). However, in my
SAS class I learned that the D should be max{|edf(1) - edf(0)|}; edf(0)
being the cumulative percent of non-responders. I was wondering if K-S can
be derived from comparing any two groups or only two mutually exclusive
groups.

Thanks

6. Test kolmogorov-smirnov

7. Kolmogorov - Smirnov Test in SAS

Hi All,

I have two variables, and I would like to compare their distributions
to each other using a Kolmogorov - Smirnov Test.

Is it possible to do that in SAS?  I can find documentation to compare
the distribution to standard distributions, but not a specified data
set.

Thanks,
Reeza