By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
458,249 Members | 1,633 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 458,249 IT Pros & Developers. It's quick & easy.

Stats algorithm: Get z-score from percentile?

P: n/a
I'm looking from an algorithm that will do a statistics calculation for me--
it's the first time 'Numerical Recipies in C' has failed me! A normal
distribution table tells me the probability (by percentile) of an item,
given something called a z-score. If I have a z-score, I can look up the
probability in a distribution table.

There are plenty of algorithms around to do that calculation, but I need to
do a reverse lookup. I am given a percentile, and I need to return the
z-score that corresponds to that percentile. I haven't found anything that
will give me that result. If anyone can direct me to code or to an
algorithm, I'll be very grateful.

Thanks,
Dave Veeneman
Nov 15 '05 #1
Share this Question
Share on Google+
3 Replies


P: n/a
What about http://www.fourmilab.ch/rpkp/experim...is/zCalc.html? Is
that what you're looking for (it's all done in JavaScript)?

Failing that, there is always Newton's Method.

Colin

"Dave Veeneman" <no****@nospam.com> wrote in message
news:%2****************@TK2MSFTNGP12.phx.gbl...
I'm looking from an algorithm that will do a statistics calculation for me-- it's the first time 'Numerical Recipies in C' has failed me! A normal
distribution table tells me the probability (by percentile) of an item,
given something called a z-score. If I have a z-score, I can look up the
probability in a distribution table.

There are plenty of algorithms around to do that calculation, but I need to do a reverse lookup. I am given a percentile, and I need to return the
z-score that corresponds to that percentile. I haven't found anything that
will give me that result. If anyone can direct me to code or to an
algorithm, I'll be very grateful.

Thanks,
Dave Veeneman

Nov 15 '05 #2

P: n/a
Thanks! I had found fourmilab, but then I assumed that the code would be
server-side. Guess it shows what happens when I assume...

I did find a Fortran version at http://lib.stat.cmu.edu/apstat/241. It's
from the Journal of Applied Statistics-- I may use that one, simply because
I can verify it in the journal.

Regards,
Dave Veeneman
Nov 15 '05 #3

P: n/a
Hey, would you be willing to share your code once you code
it up in C#? I'm trying to get together a bunch of stats
code in C#. I did a bunch in VB 6.0 for basic things
(correlation, sd, etc) and it is looking like I'll have to
convert it all.
Geoff Collier
Nov 15 '05 #4

This discussion thread is closed

Replies have been disabled for this discussion.