On Sun, 1 Jun 2008 09:16:01 -0700, Suparna
wrote:

http://office.microsoft.com/en-us/ex...118931033.aspx

http://www.cpearson.com/excel/newposte.htm

--
Kind regards,

Niek Otten
Microsoft MVP - Excel

"Suparna" wrote in message ...
|

Niek wrote on Sun, 1 Jun 2008 18:29:38 +0200:

http://www.cpearson.com/excel/newposte.htm


Possibly there may be a caution on the page but RAN(), in the
earlier versions of Excel is not very good.

--

James Silverton
Potomac, Maryland

E-mail, with obvious alterations:
not.jim.silverton.at.verizon.not

On Jun 1, 9:26*am, Lars-Åke Aspelin wrote:
http://office.microsoft.com/en-us/ex...118931033.aspx

Great pointer!

Going a little off-topic, but hopefully finding some interest among
readers of this thread....

Can anyone provide VB code that demonstrates how to generate random
data in a normal distribution with known (sampling) mean and standard
deviation __without__ relying on Excel features such as NORMINV()?

I do not want to rely on Excel features because I actually want to
translate the solution into C. (For a Unix system, not Windows.)

I did this a few years ago, but I've forgotten how.

I think I used Box-Mueller. But when I look at the algorithm today, I
do not see how it works for a specific mean and standard deviation.

Extra credit: It would be ideal if I could use a beta distribution
instead of a normal distribution. (That is, I am interested in
__both__.) But I have never understood the theory and how to control
the shape of the curve using the alpha and beta parameters. And
perhaps for that reason, I have never understood how to implement it
in C.

Again, I played with it some years ago. But I have little memory of
what I learned, if anything.

Anyway, any turnkey solutions would be appreciated. I don't have much
time to spend on this, and I have access to a Unix system for only a
short time.

TIA.

One possibility would be to use the VBA routines for inverse distributions
that are available in
http://members.aol.com/iandjmsmith/Examples.xls.

These could be translated into C, and are much more accurate than the native
Excel functions.

Jerry

" wrote:

On Jun 1, 9:26 am, Lars-Ã…ke Aspelin wrote:
http://office.microsoft.com/en-us/ex...118931033.aspx

Great pointer!

Going a little off-topic, but hopefully finding some interest among
readers of this thread....

Can anyone provide VB code that demonstrates how to generate random
data in a normal distribution with known (sampling) mean and standard
deviation __without__ relying on Excel features such as NORMINV()?

I do not want to rely on Excel features because I actually want to
translate the solution into C. (For a Unix system, not Windows.)

I did this a few years ago, but I've forgotten how.

I think I used Box-Mueller. But when I look at the algorithm today, I
do not see how it works for a specific mean and standard deviation.

Extra credit: It would be ideal if I could use a beta distribution
instead of a normal distribution. (That is, I am interested in
__both__.) But I have never understood the theory and how to control
the shape of the curve using the alpha and beta parameters. And
perhaps for that reason, I have never understood how to implement it
in C.

Again, I played with it some years ago. But I have little memory of
what I learned, if anything.

Anyway, any turnkey solutions would be appreciated. I don't have much
time to spend on this, and I have access to a Unix system for only a
short time.

TIA.

BTW, Box-Muller generates N(0,1) variates from U(0,1) variates. N(0,1)*S+m
follows the N(m,S^2) distribution.
http://en.wikipedia.org/wiki/Box-Muller_transform

Jerry

" wrote:

On Jun 1, 9:26 am, Lars-Ã…ke Aspelin wrote:
http://office.microsoft.com/en-us/ex...118931033.aspx

Great pointer!

Going a little off-topic, but hopefully finding some interest among
readers of this thread....

Can anyone provide VB code that demonstrates how to generate random
data in a normal distribution with known (sampling) mean and standard
deviation __without__ relying on Excel features such as NORMINV()?

I do not want to rely on Excel features because I actually want to
translate the solution into C. (For a Unix system, not Windows.)

I did this a few years ago, but I've forgotten how.

I think I used Box-Mueller. But when I look at the algorithm today, I
do not see how it works for a specific mean and standard deviation.

Extra credit: It would be ideal if I could use a beta distribution
instead of a normal distribution. (That is, I am interested in
__both__.) But I have never understood the theory and how to control
the shape of the curve using the alpha and beta parameters. And
perhaps for that reason, I have never understood how to implement it
in C.

Again, I played with it some years ago. But I have little memory of
what I learned, if anything.

Anyway, any turnkey solutions would be appreciated. I don't have much
time to spend on this, and I have access to a Unix system for only a
short time.

TIA.