 I don't understand the factor of "2" ; which also appears in my real problem. Here is the sample code from doc.sagemath.org/html/en/reference/quadratic_forms/sage/quadratic_forms/quadratic_form.html#sage.quadratic_forms.quadratic_form.QuadraticForm.rational_diagonal_form

Q = QuadraticForm(ZZ, 4, range(10))
D, T = Q.rational_diagonal_form(return_matrix=True)
D
[ -1/16 0 0 0 ]
[ * 4 0 0 ]
[ * * 13 0 ]
[ * * * 563/52 ]

but

T.transpose() * Q.matrix() * T
[  -1/8      0      0      0]
[     0      8      0      0]
[     0      0     26      0]
[     0      0      0 563/26]

Off by a factor of 2?

The description says.

OUTPUT: either D (if return_matrix is false) or (D,T) (if return_matrix is true) where
D – the diagonalized form of this quadratic form.
T – transformation matrix. This is such that T.transpose() * self.matrix() * T gives D.matrix().**

Ray

edit retag close merge delete

Sort by » oldest newest most voted Q.matrix() ( i.e. the .matrix() operation) returns
matrix() Returns the Hessian matrix A for which Q(X) = (1/2)∗Xt∗A∗X(1/2)∗Xt∗A∗X.
I can understand where the two comes from if your doing differentiation.
What I don't understand is why .matrix() gives this form. Ray

more