oaaclass.m

this a SVM toolbox,it is very useful for someone who just learn SVM.In order to be undestood easily,

function [labels,dfce,multi_dfce] = oaaclass(data,model)
% OAACLASS One-Against-All SVM classifier.
%  [labels] = oaaclass(data,model)
%
% Inputs:
%  data [dim x num_data] data to be classified.
%  Model [struct] classification rules.
% 
% Outputs:
%  labels [1,num_data] predicted labels.
%  dfce [1,num_data] value of the decision function.
%

% Statistical Pattern Recognition Toolbox, Vojtech Franc, Vaclav Hlavac
% (c) Czech Technical University Prague, http://cmp.felk.cvut.cz
% Modifications
% 26-aug-2002, VF
%  9-july-2002, VF


[dim,num_tst_data] = size(data);

multi_dfce=zeros(model.num_classes,num_tst_data);

for class=1:model.num_classes,

   [tmp,multi_dfce(class,:)] = svmclass2(data,model.trn_data,...
       multi2dicho(model.trn_labels,class),...
       model.rule{class}.Alpha,model.rule{class}.bias,...
       model.SVM.kernel,model.SVM.arg);
end

[dfce,labels]=max(multi_dfce);

return;

%EOF