In article <[EMAIL PROTECTED]>,
Douglas C. <[EMAIL PROTECTED]> wrote:
>Suppose a rv has support only on R+ (i.e. a non-negative rv). Given
>integer moments of x, m1, m2,...mn, how might I compute bounds on some
>_other_ moment m_n+1?

>For instance, suppose I fix the first and second moment. What are the
>bounds on the third moment?

>If this is fully covered in some text, I'd appreciate a reference. 

This is fully covered in Shohat and Tamarkin, and probably in 
any other book on the moment problem.  

The conditions are that the expected value of a non-negative
polynomial (or function, if other than consecutive powers are
used) is non-negative.  For the n+1 moment on R+, this means
that the appropriate one of the matrices with a_ij = m_i+i 
or b_ij = m_i+j+1, i, j starting at 0, with m_n+1 the last
element, is positive semi-definite, assuming no previous one
is singular.  One singular matrix determines the distribution exactly.
-- 
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
[EMAIL PROTECTED]         Phone: (765)494-6054   FAX: (765)494-0558
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to