svm_vccore.m

最新的模式识别分类工具箱,希望对朋友们有用!

% Learns classifier and classifies test set
%  using the perceptron learning algorithm
%  Works with 2 class labels, any number of features
%  when the class labels are -1 and 1.
%  Invoke using SVM_VC
%  
% Inputs:
% Usage
%      [trainError, testError, estTrainLabels, estTestLabels] = ...
%           SVM_VCcore(trainFeatures, trainLabels,algParam ,testFeatures, testLabels)
% where
%
% Inputs:
% 	trainFeatures  - the training set vectors, one vector per column
%	trainLabels    - the labels of the above
%       algParam       - algorithm parameters:
%                        kernel     + kernelparameters
%                        'linear'       None
%                        'poly'         polynomial order
%                        'rbf'          scale factor
%                        Note: kernel 'rbffull' is not used
%                        
%       testFeatures   - test set, one column per vector
%       testLabels     - labels  for test set%
% Outputs
%	trainError     - the error rate on the training set (one entry per
%	                   class + total error)
%	testError      - the error rate on the test set (one entry per class
%	                  + total error)
%       estTrainLabels - the labels produced by the algorithm for the
%                          training samples
%       estTestLabels - the labels produced by the algorithm for the
%                          test samples

function   [trainError, testError, estTrainLabels, estTestLabels] = ...
    SVM_VCcore(trainFeatures, trainLabels,algParam,testFeatures, testLabels)


comma_loc = findstr(algParam,',');
kernel    = algParam(3:comma_loc(1)-2);
kernelpar = str2num(algParam(comma_loc(1)+1:length(algParam)-1));

[Nclasses, classes] = find_classes([trainLabels(:);testLabels(:)]);

if ( max(classes) == min(classes))
  estTrainLabels = ones(size(trainLabels)) * max(classes);
  estTestLabels  = ones(size(testLabels))  * max(classes);
elseif (max(classes) ~= 1 | min(classes) ~= -1),
  if ( min(classes)==0)
    trainLabels(find(trainLabels==0))=-1;
    testLabels(find(testLabels==0))=-1;
  else
    fprintf('When invoking this classifier, use only labels -1 and 1\n');
    return;
  end
end

trainFeatures = trainFeatures'; 
% this is how the algorithms appears to work;
[Nsam, Dim] = size(trainFeatures);

net = construct_svm(Dim,kernel,kernelpar,100); % default value
net = svmtrain(net,trainFeatures,trainLabels',[]);

testFeatures = testFeatures';

[estTrainLabels,foo] = svmfwd(net,trainFeatures);
[estTestLabels,foo]  = svmfwd(net,testFeatures);

trainError = computeError( classes, trainLabels, estTrainLabels);
testError  = computeError( classes, testLabels, estTestLabels);