KING

The method in the system

• MultiNomial logistic Regressoin
  • bad in high dimension
    • Factor = mnrfit(train_data, train_label);
    • scores = mnrval(Factor, test_data);

• Random Forest

  • good classifier
  • score is the probability output
    • Factor = TreeBagger(numberoftree, train_data, train_label);
    • [predict_label, scores] = predict(Factor, test_data);

• Navie Bayes

  • Factor = NaiveBayes(train_data, train_label);
  • Scores = posterior(Factor, test_data);
  • [Scores, predict_label] = posterior(Factor, test_data);
  • predict_label = predict(Factor, test_data);
  • accuracy = length(find(predict_label == test_label))/length(test_label)*100;
• SVM
  • Factor = svmtrain(train_label, train_data, '-b 1');
  • [predict_label, accuracy, scores] = svmpredict(test_label, test_data, Factor, '-b 1');
• KNN
  • predict_label = knnclassify(test_data, train_data, train_label, num_neighbors);
  • In matlab 2012
    • Factor = ClassificationKNN.fit(train_data, train_label, 'NumNeighbors', numofneighbours);
    • [predict_label, scores] = predict(Factor, test_data);
• Ensembles for boosting, Bagging or Random Subspace
  • In matlab2012
    • Factor = fitensemble(train_data, train_label, 'AdaBoostM2', 100, 'tree');
    • Factor = fitensemble(train_data, train_label, 'AdaBoostM2', 100, 'tree', 'type', 'classification');
    • Factor = fitensemble(train_data, train_label, 'SubSpace', 50, 'KNN');
    • [predict_label, scores] = predict(Factor, test_data);
• discriminant analysis classifier
  • Factor = ClassificationDiscriminant(train_data, train_label);
  • Factor = ClassificationDiscriminant(train_data, train_label, 'discrimType', 'non-linear...');
  • [predict_label, scores] = predict(Factor, test_data);