fits spin wave spectra to experimental spectral data
fitsp = FITSPEC(obj, 'Option1', Value1, ...)
Options:
epsilon Small number that controls wether the magnetic structure is
incommensurate or commensurate, default value is 1e-5.
func Function to change the Hamiltonian in obj, it has the following
header:
obj = @func(obj, x);
datapath Path to the file that stores the experimental data. For the
input data format see <a href="matlab:doc sw_readspec">sw_readspec</a>.
Evect Vector, defines the energy binning of the calculated
dispersion. Larger binning steps solve the issue of fitting
unresolved modes. Size is [1 nE].
xmin Minimum limit of the optimisation parameters, optional.
xmax Maximum limit of the optimisation parameters, optional.
x0 Starting value of the optimisation parameters. If empty
or undefined, then random values are used.
nRun Number of consecutive fitting runs, each result is saved in the
output fitsp.x and R arrays. If the Hamiltonian given by the
random x parameters is incompatible with the ground state,
those x values will be skipped and new random x values will be
generated. Default is 1.
nMax Maximum number of runs, including the ones that produce error
(due to incompatible ground state). Default is 1000.
hermit Method for matrix diagonalization:
true J.H.P. Colpa, Physica 93A (1978) 327,
false R.M. White, PR 139 (1965) A450.
Colpa: the grand dynamical matrix is converted into another
Hermitian matrix, that will give the real eigenvalues.
White: the non-Hermitian g*H matrix will be diagonalised,
that is not the elegant method.
Advise:
Always use Colpa's method that is faster, except when small
imaginary eigenvalues are expected. In this case only White's
method work.
Default is true.
Parameters for visualizing the fit results:
plot If true, the measured dispersion is plotted together with the
fit. Default is true.
wFact Weigth of the spin wave enenergies, plotted as the width of the
data points (vertical lines).
iFact Factor of the plotted simulated spin wave intensity (red
ellipsoids).
lShift Vertical shift of the Q point labels on the plot.
Optimisation options:
tolx Minimum change of x when convergence reached, default
value is 1e-4.
tolfun Minimum change of the R value when convergence reached,
default value is 1e-5.
maxfunevals Maximum number of function evaluations, default value is
1e7.
Output:
Output fitsp is struct type with the following fields:
obj Copy of the input sw class object, with the best fitted
Hamiltonian.
x Final values of the fitted parameters, dimensions are
[nRun nPar]. The rows of x are sorted according to increasing R
values.
R R-value, goodness of the fit, dimensions are [nRun 1], sorted
in increasing order.
exitflag Exit flag of the fminsearch command.
output Output of the fminsearch command.
Any option used by SW.SPINWAVE function are also accepted.
See also SW.SPINWAVE, SW_EGRID, SW_NEUTRON, SW_READSPEC, FMINSEARCH.