next up previous contents
Next: Independence property Up: Information correlationcumulants, clusters, Previous: Cumulants in one dimension

 

Cumulants in many dimensions

In the case of many random variables we have a similar generating function giving the various moments. tex2html_wrap_inline13227, Here we have
equation847
The cumulant expansion in this case is then given by equating coefficients of the tex2html_wrap_inline13229 in
 equation1219
where tex2html_wrap_inline13235 indicates tex2html_wrap_inline13237, and tex2html_wrap_inline13239 in tex2html_wrap_inline13241 consists of tex2html_wrap_inline13243 tex2html_wrap_inline13245 indices, etc. The first four cumulants are [27]
  eqnarray883
In general, all distinct terms where each variable appears exactly once are included, and the coefficient of any term is tex2html_wrap_inline13247 where n is the number of averages in the term.



David Wolf
Tue Mar 25 08:11:49 CST 1997