Matrixexponential and Normaldistribution/Gauss' function (II)
Gottfried Helms
Annette.Warlich at t-online.de
Sat Mar 4 23:24:35 CET 2006
Sorry, I was a bit sloppy with the copy&paste of results.
I have to correct two displays:
Am 04.03.2006 21:54 schrieb Gottfried Helms:
> While the matrix-logarith of the binomial-triangle has the sequence
> of natural numbers n in the *first* subdiagonal, we have here the triangular
> numbers (2n-1)*2n / 2 in the *second* subdiagonal (but only each second
> entry is different from zero.
>
> I would call things like that maximal compression of information...
> ... and a strong surprise.
>
> But one can even smooth that thing and proceed;
> use A as the "Gauss-triangle", and compute its square:
>
> A1 = A*A
> | 1 0 0 0 0 0 0 0 |
> | 0 1 0 0 0 0 0 0 |
> | -2 0 1 0 0 0 0 0 |
> | 0 -6 0 1 0 0 0 0 |
> | 12 0 -12 0 1 0 0 0 |
> | 0 60 0 -20 0 1 0 0 |
> | -120 0 180 0 -30 0 1 0 |
> | 0 -840 0 420 0 -42 0 1 |
here I took the wrong output; the numbers must be neagtive;
> and now the matrix logarithm has no holes in the sequence:
> lA1 = ln(A*A)
> | 0 0 0 0 0 0 0 0 |
> | 0 0 0 0 0 0 0 0 |
> | 2 0 0 0 0 0 0 0 |
> | 0 6 0 0 0 0 0 0 |
> | 0 0 12 0 0 0 0 0 |
> | 0 0 0 20 0 0 0 0 |
> | 0 0 0 0 30 0 0 0 |
> | 0 0 0 0 0 42 0 0 |
that means, I forgot to document the inversion-step; it is
lA1 = ln(inv(A*A)) or lA1 = - ln(A*A)
and as well with
>
> then the matrix-logarithm has the -say- most simple representation (second only
> to that for the binomial-matrix):
>
> LA2 = ln(A2)
should be LA2 = - ln(A2) or LA2 = - ln(sqrt(A*A)) or LA2 = ln( inv (sqrt(A*A)))
> | 0 0 0 0 0 0 0 0 |
> | 0 0 0 0 0 0 0 0 |
> | 1 0 0 0 0 0 0 0 |
> | 0 3 0 0 0 0 0 0 |
> | 0 0 6 0 0 0 0 0 |
> | 0 0 0 10 0 0 0 0 |
> | 0 0 0 0 15 0 0 0 |
> | 0 0 0 0 0 21 0 0 |
>
and then in the second eq I missed the matrix-inversion, sorry: it must be
>
> exp("natural numbers") = "binomial-triangle"
> _______________________
> exp("triangular-numbers") = \/inv("Gauß-triangle") ²
which does not affect the numbers, only their sign -
but obviously it loses much of its simplicticy. One can repair that by
> _________________
> exp(- "triangular-numbers") = \/"Gauß-triangle" ²
it's a pity...
But anyway, I feel the last note is still valid:
>
> This seems to be the start of an interesting family of numbers/functions...
>
Gottfried
More information about the SeqFan
mailing list