Talk:Conference matrix

Can anyone add to the introduction more information about what conference matrices are used for, and how? Zaslav 04:10, 15 November 2006 (UTC)[reply]

The intro mentions some uses of conference matrices with references but wouldn't it be better to have a short section that explains uses? If someone has time to do it, that would be nice! Zaslav (talk) 10:41, 16 March 2008 (UTC)[reply]

I have reversed the bold-facing of the matrix names, which seems to be due to a misunderstanding. The "C" in C-matrix is not a matrix, it is the abbreviation of the word "conference". Zaslav (talk) 04:21, 10 January 2008 (UTC)[reply]

position of the 0

Obviously, the condition to have exactly one 0 on each row and column is a consequence of CC'=(n-1)I, (since diagonal elements are sum( Cij^2 , j=1..n) for any given i) but there are research papers that do not require the 0 to be on the diagonal, e.g. http://dx.doi.org/10.1016/j.jcta.2005.05.005 where we read:

A conference matrix C of order n is a square (0,±1) matrix of side n with exactly one 0 in every row and every column such that CC' = (n−1)I , where I is the identity matrix and C' denotes the transpose of C.

or http://dx.doi.org/10.1016/j.endm.2004.03.036 where we read

Definition 3.1 A weighing matrix W(n,w) of weight w!=0 and order n is a square matrix of size n with entries from {−1, 0, +1} satisfying WW^t = wI. A W(n, n) is called an Hadamard matrix of order n, a W(n, n−1) is called a conference matrix of order n.

A possibility of such a matrix which does not have the 0 on the diagonal is:

[1,  0,  1,  1]
[0, -1, -1,  1]
[1, -1,  0, -1]
[1,  1, -1,  0]

If both definitions are used, it should be mentioned. — MFH:Talk 14:39, 14 March 2008 (UTC)[reply]