train_test_validate.m

一个matlab的工具包,里面包括一些分类器 例如 KNN KMEAN SVM NETLAB 等等有很多.

% Input pararmeter: 
% D: data array, including the feature data and output class
% outputfile: the output file name of classifiers
function run = train_test_validate(D, classifier_wrapper_handle, classifier)

global preprocess;
[X, Y, num_data, num_feature] = preprocessing(D);
clear D;

% The statistics of dataset
% num_class = length(preprocess.ClassSet);
% actual_num_class = length(preprocess.OrgClassSet);
% num_shot = length(preprocess.ShotIDSet);
% class_set = preprocess.ClassSet;

if (preprocess.TrainTestSplitBoundary > 0),
    splitboundary = preprocess.TrainTestSplitBoundary;
else
    splitboundary = fix(num_data / (-preprocess.TrainTestSplitBoundary));
end;
testindex = splitboundary+1:num_data;
trainindex = 1:splitboundary;

% trainindex = []; testindex = [];
% for i = 1:length(preprocess.OrgClassSet),
%     ind = find(Y == preprocess.OrgClassSet(i));
%     sb = fix(2*length(ind)/3);
%     trainindex = [trainindex; ind(1:sb)];
%     testindex = [testindex; ind(sb+1:length(ind))];    
% end;

run = feval(classifier_wrapper_handle, X, Y, trainindex, testindex, classifier); 
  
run.Y_pred = zeros(length(testindex), 4);
run.Y_pred(:, 1) = (1:length(testindex))';
run.Y_pred(:, 2) = run.Y_prob; 
run.Y_pred(:, 3) = run.Y_compute; 
run.Y_pred(:, 4) = run.Y_test;