[gmx-users] Derivative of eigenvectors

[Ghermes Chilov]

Hi!

D> On Thu, 2003-02-27 at 18:26, Nguyen Hoang Phuong wrote:

>> I would like to compute the time derivative of the eigenvectors and
>> eigenvalues of the covariance matrix. Does anyone know the methods?
D> No.. the elements of the covar matrix are averages over time, so what
D> would be the meaning of that?

this would be the increment of the mean-value (matrix element or an
eigenvector) in time. Of course it will depend on the time interval
(window length) chosen for averaging, but i guess it could have sense
in analyzing transitions between strongly coupled states: within each
state the derivative vector is zeroth, but it undergoes break
(disruptious behavior) between them.

why not?!

Ghermes Chilov
PhD student, Moscow State University