Acceptance Rejection Sampling

[Exp i3Pi]

Hello,

The wikipedia article details the following. I dont understand certain
things with it as I am new to statistics. Please explain me how the
following equation is written in the article http://en.wikipedia.org/wiki/Rejection_sampling

Pr( u < f(x)/(M*g(x)) ) = E[f(x)/(M*g(x))]

how did we get the expected value from the probability. It doesn't
make much sense.

Thanks,
Ramesh

[Herman Rubin]

In article <4ee5c9be-724c-46d1...@a31g2000yqn.googlegroups.com>,

It is the other way. Probability is the expected value
of conditional probability, so

Pr( u < f(x)/(M*g(x)) = E(Pr( u < f(x)/(M*g(x)| x)
= E[f(x)/(M*g(x))]

since P(u < c) = c if 0 <= c <= 1.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hru...@stat.purdue.edu Phone: (765)494-6054 FAX: (765)494-0558