Next: Method of calculations
Up: Theoretical framework
Previous: Additivity of effective s.p.
Determination of effective s.p. observables
Once the averages of physical observables for the set of
calculated configurations () are determined, the
effective s.p. contributions
(19) are found by means of a multivariate least-square fit
(see, e.g., Refs. [29,30]). This is done by minimizing
the function of
defined by
|
(22) |
Note that the problem is only meaningful when the number of configurations
is sufficiently large, . Following
the general concept of the least-square method,
the partial differentiation with respect to the variables
yields
where
and
. Solving this equation by inverting the
non-singular matrix gives the solution to the multivariate regression problem:
|
(24) |
The fact that is positive-definite guarantees that the solution
corresponds to a minimum of
.
In order to estimate the
variance, we assume that the first statistical moments of residuals,
|
(25) |
are zero for all . Consequently,
can be
considered an unbiased estimate of
. Furthermore,
under the assumption that
|
(26) |
for all
, and
|
(27) |
for all
, one can define the
variance-covariance matrix as
, for which the unbiased estimate for is given
by
|
(28) |
Finally, the unbiased estimate for the variance-covariance matrix for
is given by
. In what follows
we do not differentiate between notations for variables and their
estimates.
The least-square procedure described in this section was used to
determine the effective s.p. quadrupole moments
and angular momentum alignments
.
Next: Method of calculations
Up: Theoretical framework
Previous: Additivity of effective s.p.
Jacek Dobaczewski
2007-08-08