home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cis.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!ucbvax!stat!s135
- From: s135@stat.Berkeley.EDU (Chad Heilig)
- Newsgroups: sci.math.stat
- Subject: Geometric Standard Deviation
- Message-ID: <44651@ucbvax.BERKELEY.EDU>
- Date: 25 Aug 92 21:55:45 GMT
- Sender: nobody@ucbvax.BERKELEY.EDU
- Organization: Statistics Dept., U. C. Berkeley
- Lines: 27
-
- Keywords: geometric standard deviation
-
- I need to find a reference to verify the definition of the geometric
- standard deviation. Here is the particular problem:
-
- Given a set of numbers x_1,...,x_n, let y_i = log x_i.
- It can be shown that the geometric mean of the x_i is the same as
- exp(ybar), where ybar is the arithmetic mean of the y_i.
-
- I want to confirm that the geometric SD of the x_i is also
- exp(ysigmahat), where ysigmahat is the standard deviation of the y_i.
-
- I have a hunch that this is the way geometric SD is defined, but
- I'd happily yield to corrections. Moreover, if I am wrong, I'd
- like to find a link between the GSD and exp(ysimgahat).
-
- I am using this information to justify one way of estimating
- the GM and GSD of a lognormal distribution without having to
- explain MLE's and MME's.
-
- Please reply by e-mail to this account; I'll summarize the responses
- and re-post later.
-
- Thanks.
-
- Chad
- s135@stat.Berkeley.EDU
-