Re: Lilliefors Test : 40 years
- From: hrubin@xxxxxxxxxxxxxxxxxxxx (Herman Rubin)
- Date: 28 Apr 2008 13:34:43 -0400
In article <22560633.1209402134324.JavaMail.jakarta@xxxxxxxxxxxxxxxxxxxxxx>,
Luis A. Afonso <licas_@xxxxxxxxxxx> wrote:
All Readers agree that the so- called Kolmogorov Test IS ABSOLUTELY WORHLESS by the only reason that one MUST know the parameters exact values (mean and variance), WHICH IS AN ABHERATION IN STATISTICAL TERMS, in order to perform the calculations or to read the critical values on Tables,
Luis Amaral Afonso (The Moderator Destroyer)
The Kolmogorov-Smirnov test is a good test if the correct
distribution is used. If the population distribution is
completely specified, there is no problem. If it is not,
the distribution depends on the way parameters are
estimated, translated appropriately.
The distribution can be simulated, and for small samples,
it seems unlikely anything else can be done. In principle,
the asymptotic distribution can be simulated as well, but
be careful; the Brownian bridge, with linear functionals
set equal to zero, can be simulated pointwise, but it will
be necessary to simulate the maxima and minima between the
points, as this author did in studies of the comparison of
tests for shift in location parameter whose asymptotic
properties (for VERY large samples) were known to be
equivalent.
Herman Rubin, presenting the evidence
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@xxxxxxxxxxxxxxx Phone: (765)494-6054 FAX: (765)494-0558
.
- References:
- Re: Lilliefors Test : 40 years
- From: Luis A. Afonso
- Re: Lilliefors Test : 40 years
- Prev by Date: Re: Lilliefors Test : 40 years
- Next by Date: Why we created ?
- Previous by thread: Re: Lilliefors Test : 40 years
- Next by thread: Re: Lilliefors Test : 40 years
- Index(es):
Relevant Pages
|
|
