KFD |
Kernel Fisher Discriminat.
Synopsis:
model = kfd( data )
model = kfd( data, options )
Description:
This function is an implementation of the Kernel Fisher
Discriminant (KFD) [Mika99a]. The aim is to find a binary
kernel classifier which is the linear decision function in a
feature space induced by the selected kernel function.
The bias is found decision function is trainined by the
linear SVM on the data projected on the optimal direction.
Input:
data [struct] Training binary labeled data:
.X [dim x num_data] Vectors.
.y [1 x num_data] Labels (1 or 2).
options [struct] Control parameters:
.ker [string] Kernel identifier (default 'linear').
See 'help kernel' for more info.
.arg [1 x nargs] Kernel argument(s).
.C [1x1] Regularization constant of the linear 1-D SVM
used to optimize the bias (default C=inf).
.mu [1x1] Regularization constant added to the diagonal of
the within scatter matrix (default 1e-4).
Output:
model [struct] Binary SVM classifier:
.Alpha [num_data x 1] Weight vector.
.b [1x1] Bias of decision function.
.sv.X [dim x num_data] Training data (support vectors).
.trnerr [1x1] Training classification error.
.kercnt [1x1] Number of kernel evaluations used during training.
.nsv [1x1] Number of support vectors.
.options [struct] Copy of options.
.cputime [1x1] Used cputime.
Example:
trn = load('riply_trn');
options = struct('ker','rbf','arg',1,'C',10,'mu',0.001);
model = kfd(trn, options)
figure; ppatterns(trn); psvm(model);
See also
SVMCLASS, FLD, SVM.
Modifications:
17-may-2004, VF
14-may-2004, VF
7-july-2003, VF