Re: Help: Non-positive semi-definite message

From: Alan Xiao <Alan.Xiao@cognigencorp.com> Subject: Re: Help: Non-positive semi-definite message Date: Mon, 20 Aug 2001 17:37:55 -0400 Thanks, Ken, I noticed that you and Diane and others talked about this online months ago. Yee, this is a kind of playing trick, you reduce a parameter eta but increase a parameter theta - the total parameter numbers are the same (Of course they have different meanings) and you get the same (?) (other) parameter estimates and the porblem is gone. How about some kind of correlation between ETAs but not equal (close) to 1 (-1)? Does THETA(3) in your example have to be positive? And what's the difference between using this method and $OMEGA BLOCK? If any of you has already tested these kinds of quesitons and their inferences are general, and you are willing to share them with me (and other NMusers) here, that would be great - and I really really appreciate that. My time frame does allow me to test (even think) every aspect for every possible solution. Thanks again, Alan.