Dear All,

I'm working at a manufacture as Statistical Process Control officer.

I'm looking for references about CpK when the process is directed to the lower spec.
As we know, the CpK formulation is:

MIN (((USL-Xbar)/(3*(Rbar/d2))),((X-bar-LSL)/(3*(Xbar/d2)))

Because our process is directed to the lower spec, the average will getting lower and the CpK value will getting lower too. That's why it's not easy for us to meet the CpK target.

If any of you have a reference to cover this problem, please share...

Thank You,

Regards,
- Irma -
Bintan Island - Indonesia

Bonjour a tous,
j'ai 3 parametres et 5 class et les mesures sont repetees. Je cherche
un model de regression pour identifier les parametres significatifs.
Dans le cas de 2 class(binaire) on applique (GEE), mais le cas de plus
que 2 (dans mon cas 5), je ne sais pas comment faire, pouvez vous
m'aider a resoudre ce probleme, et si quelqu'un sait comment faire le
code SAS, je le prie qu'il m'envoie ce code.
Merci beaucoup

 




________________________________________________________________________
NOTE: The Statisticians_group is for discussing on different issues related to the subject Statistics, Statistical softwares like SAS, SPSS etc. The users are encouraged to post jobs those are exclusively for Statisticians like data analyst, business analyst etc.

Please don't post or send any SPAM!

________________________________________________________________________

Bonjour a tous,
j'ai 3 parametres et 5 class et les mesures sont repetees. Je cherche
un model de regression pour identifier les parametres significatifs.
Dans le cas de 2 class(binaire) on applique (GEE), mais le cas de plus
que 2 (dans mon cas 5), je ne sais pas comment faire, pouvez vous
m'aider a resoudre ce probleme, et si quelqu'un sait comment faire le
code SAS, je le prie qu'il m'envoie ce code.
Merci beaucoup

 

---

Yahoo! Groups Links

• To visit your group on the web, go to:
  http://in.groups.yahoo.com/group/Statisticians_group/
   
• To unsubscribe from this group, send an email to:
  Statisticians_group-unsubscribe@...
   
• Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.

Hi Im a student taking statistics. I really need help immediately with
this problem:

The following hypothesis are given:

Ho: p is less or equal to 0
H1: p is greater than 0

   A random sample of 12 observations indicated a correlation of .32
Can we conclude that the correlation in the population is greater than 0?
Use the .05 significance level.

Ho is rejected if t>1.714

t= .32SQRt 12-5/SQRT 1-(32)^2=?

    I don't know what to do after this. Would someone please help me?

I am in dire need for help with some statistics work.

Hello friends!

 

I would like to know what mistakes we get when we weight a sample to have the same distribuition of the population. 

 

For example, if the population has:

 

Group A: 20%

Group B: 40%

Group C: 30%

Group D: 10%

 

And the sample has:

 

A: 30

B: 10

C: 25

D: 35

Total: 100

 

How this impact the study? It is confiable, if we weight the sample to have the same distribuition of the population?

How do we know if the 10 units of group B represent the 40% of the population?

 

I dont know if my question is clear, but... 

 

Thanks, anyway!!

 

Regards

Rosana

 

Abraços,

Rosana

 

Hi,

Anybody has any experience in getting the separability between image
classes using the satatistical method 'transformed divergence'?
If anybody interested I can post more details so we can discuss this
issue...

Any comments would be highly appreciated

Bill

hi..im new in field of banking..im just hired 3 months ago..i would
like to ask to those experienced statisticians who worked in
bank...what statistical analysis are usually used in banks.

Hi Dear!

Please let me know about SAS versions. (i.e. base version, advance etc)

Thank You.

Hi Dear!

Please let me know about SAS versions. (i.e. base version, advance etc)

Thank You.

hello
i am not a statistican but i am electrical engineering and doing
master in electrical engineering from tu-darmstadt germany i am doing
specilization in signal procesiing as all know that signal procesing
requires alot to have knowledge from stochastic randon variables
(Terms like mean,variance,corelation erogodicity......and many more )
what i am now looking from people here that give me some link from i
could find an easy to undrstand things to have very good basics in
there topics
regards

Hi,
   I am trying to understand application of sub-space
method for outlier detection.
   If anyone has any good pointer to understand the
theory, I will appreciate that since the following
  good pointer will hard to parse:

   http://www3.interscience.wiley.com/cgi-bin/abstract/108067748/ABSTRACT

Thanks,
Dhman

Hi,

Anybody has any experience in getting the separability between image
classes using the satatistical method 'transformed divergence'?
If anybody interested I can post more details so we can discuss this
issue...

Any comments would be highly appreciated

Bill

Madan Gopal Kundu wrote:
> Hi all,
>
> Whhat is the logic behind plotting residuals against expected value of
> dependent variable in residual analysis (for independence of error
> testing). Why not residuals against observed value of dependent
> variable ?
>

Think about it this way:

A diagnostic plot like this can be looked at as a
(graphical) hypothesis test.  As a numerical hypothesis test should have
a well-defined (expected) null value, a graphical hypothesis test should
have a well-defined null pattern, and you should know what kind of
deviations from this to expect if the null is untrue. (a good idea is to
make a simulated null plot to compare with the actual plot).

The plot of observed versus residuals does'nt behave as you expect under
the null. Look at this simulation, in R: (www.Rproject.org)

  > x <- 1:20
  > y <- 3 + 0.5 * x + rnorm(20, sd=3)
  > plot(x,y)
Waiting to confirm page change...
  > mod <- lm( y ~ x )
  > plot( fitted(mod), resid(mod))
  > plot( y,resid(mod))
Waiting to confirm page change...
  > cor( fitted(mod), resid(mod))
[1] -4.558654e-18
  > cor( y,resid(mod))
[1] 0.7906585
  >

Look at that last correlation:

Kjetil halvorsen




> Any suuggestion will be appreciated.
>
> with regards,
>
> madan
>
>
>
>
>
> ________________________________________________________________________
> NOTE: The Statisticians_group is for discussing on different issues related to
the subject Statistics, Statistical softwares like SAS, SPSS etc. The users are
encouraged to post jobs those are exclusively for Statisticians like data
analyst, business analyst etc.
>
> Please don't post or send any SPAM!
>
> ________________________________________________________________________
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>

Hi all,

Whhat is the logic behind plotting residuals against expected value of
dependent variable in residual analysis (for independence of error
testing). Why not residuals against observed value of dependent
variable ?

Any suuggestion will be appreciated.

with regards,

madan