RE: FW: OMEGA HAS A NONZERO BLOCK

From:Leonid Gibiansky Subject:RE: FW: [NMusers] OMEGA HAS A NONZERO BLOCK Date:Fri, 04 Oct 2002 11:14:38 -0400 Ken, I was thinking about a little more general situation where only the determinant of OMEGA is equal to zero and missed additional corr(1,4)=1 information that is responsible for the other two parameters. In Steve's case the approach that you suggested should work perfectly (minor difference of correlation from 1 should not change the result). Leonid Correlation matrix: X1 X2 X3 X4 1 1.0000000 0.6852854 0.3190490 0.9984821 2 0.6852854 1.0000000 0.4729386 0.6443795 3 0.3190490 0.4729386 1.0000000 0.3676685 4 0.9984821 0.6443795 0.3676685 1.0000000