seeking reference
- From: mccolm@xxxxxxxxxxx
- Date: 25 Jun 2007 21:52:14 -0400
I am looking for a book published at least two years ago, but which is
hiding from me. So I would like to know if the following probability
(concentration of measure nomenclature) rings any bells.
There is a generalization of the notion of the median, usually
called a "quantile", but in this book called a "median." It is sort
of an inverse of the cumulative distribution function (PDF): if X is
a random variable, F_X (x) = Prob (X <= x). The inverse of this seems
to have no standard notation, but in this book it is called an e-
median (instead of a quantile), and defined: M_e [X] = inf {x : F_X
(x) > = e}. Notice that this is a perfectly good statistic (like a
variance or an expectation). The book goes on to define the "e-width"
of X as W_e [X] = M_1-e/2 [X] - M_e/2 [X].
Does this ring any bells? Any suggestions where I might look?
.
- Prev by Date: Re: al matrices
- Next by Date: Re: finding an article
- Previous by thread: finding an article
- Next by thread: nonhereditary module
- Index(es):
Relevant Pages
|
