svm_vc.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 Support Vector Machines.
% The actual SVM Code is GNU code (see SVMReadme.txt)
%
%
%  If there are 2 classes, there is no problem (except that the labels are
%     converted to 0 and 1).
%  If there are more than 2 classes, pairwise classification is made,
%  followed by majority vote.
%  
% Inputs:
% Usage
%      [trainError, testError, estTrainLabels, estTestLabels] = ...
%           SVM_VC(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'       unused number
%                        'poly'         polynomial order
%                        'rbf'          scale factor
%	                  
%       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)

function   [trainError, testError] = ...
    SVM_VC(trainFeatures, trainLabels,algoParams,testFeatures, testLabels)

[Nclasses, classes]  = find_classes([trainLabels(:);testLabels(:)]);
 % Number of classes in labels

if (Nclasses > 2)
  [trainError, testError] = classifierWrapper(trainFeatures, trainLabels, ...
      'SVM_VCcore', algoParams, testFeatures, testLabels);
else
  ind0tr = find(trainLabels == min(classes));
  ind0te = find(testLabels  == min(classes));
  ind1tr = find(trainLabels == max(classes));
  ind1te = find(testLabels  == max(classes));
  trainLabels(ind0tr) = -1;
  trainLabels(ind1tr) = 1;
  testLabels(ind0te)  = -1;
  testLabels(ind1te)  = 1;
  [trainError, testError] = ...
      SVM_VCcore(trainFeatures, trainLabels, algoParams, testFeatures, testLabels);
end


trainError = reshape(trainError,length(trainError),1);
testError  = reshape(testError, length(testError),1);