Message
Let's
say you are taking multiple independent identically distributed values from a
gaussian random value generator. The mean of those values is itself a
random variable. Say you take values X1, X2, ..., XN from a random number
generator, then the mean of them is MX = sum(X1,...,XN)/N, which itself is a
random variable. The expectation of MX (or its mean, if you formed many
MX's from different sets of Xi's, i from 1 to N) is zero, but the variance of
MX, E{ MX^2 } if MX is zero-mean, is NOT zero.
If the
variance of the random value generator was 1, then the variance of MX is
1/N. So any finite set of X's has a mean that on average will be zero, but
is not guaranteed (indeed, has infinitesimal probability of being
exactly) zero. The larger the set, the closer MX should be to
zero.
Most
likely, the number generator is working perfectly, and you are observing an
effect caused by a finite set of samples.
Sincerely,
Glen
Ragan
----Original Message----- From: mohan prasad
[mailto:m...@yahoo.com] Sent: Tuesday, March 04, 2003 1:48 PMTo: m...@yahoogroups.comSubject: [matlab] zero mean random
variablesHi,Am trying to generate complex gaussian random variables using the randn function but when I calculated the mean of those random variables it is not giving zero as answer.Can anyone tell why this is happening? mohan _____________________________________ Note: If you do a simple "reply" with your email client, only the author of this message will receive your answer. You need to do a "reply all" if you want your answer to be distributed to the entire group. _____________________________________ About this discussion group: To Join: m...@yahoogroups.com To Post: m...@yahoogroups.com To Leave: m...@yahoogroups.com Archives: http://www.yahoogroups.com/group/matlab More DSP-Related Groups: http://www.dsprelated.com/groups.php3 ">Yahoo!
Terms of Service. |