"Miller, Raul D" <[EMAIL PROTECTED]> writes: > However: in the general case it is costly to find all eigenvectors.
That I believe. In this case, I seem to have gotten all 4 quickly, so I figured there shouldn't be much advantage in using another method just to find the principal eigenvector unless there is a rather large round-off error (I do understand that eigenvalue problems can be subject to such problems). Bill -- Bill Harris http://facilitatedsystems.com/weblog/ Facilitated Systems Everett, WA 98208 USA http://facilitatedsystems.com/ phone: +1 425 337-5541 ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
