1.前言

目前瞭解到的MATLAB分類器有：K近鄰，隨機森林，樸素貝葉斯，整合學習方法，鑑別分析，支援向量機。現將其主要函式使用方法總結如下，更多細節需參考MATLAB 幫助檔案。
設：

• 訓練樣本：train_data            % 矩陣，每行一個樣本，每列一個特徵
• 訓練樣本標籤：train_label       % 列向量
• 測試樣本：test_data
• 測試樣本標籤：test_label

2.分類器使用方法

2.1 K近鄰分類器—KNN

mdl = ClassificationKNN.fit(train_data,train_label,'NumNeighbors',1);
predict_label= predict(mdl, test_data);
accuracy     = length(find(predict_label == test_label))/length(test_label)*100;

2.2 隨機森林分類器—RandomF

nTree = 20;
B = TreeBagger(nTree,train_data_pca,train_label');
predict_label = predict(B,test_data_pca);
predict_label = str2double(predict_label);
accuracy = length(find(predict_label == test_label'))/length(test_label)*100;

2.3 樸素貝葉斯分類器—Naive Bayes

nb = NaiveBayes.fit(train_data, train_label);
predict_label = predict(nb, test_data);
accuracy = length(find(predict_label == test_label))/length(test_label)*100;

2.4 整合方法分類器—AdaBoost

ens = fitensemble(train_data,train_label,'AdaBoostM1',100,'tree','type','classification');
predict_label = predict(ens, test_data);
accuracy = length(find(predict_label == test_label))/length(test_label)*100;

2.5 鑑別分析分類器—Discriminant Analysis Classifier

obj = ClassificationDiscriminant.fit(train_data, train_label);
predict_label = predict(obj, test_data);

2.6 支援向量機分類器—SVM

option = statset('MaxIter',1000000);
svm_struct = svmtrain(train_data_pca,train_label,'options',option);
predict = svmclassify(svm_struct,test_data_pca);
correct_num = sum((predict-test_label')==0);
accuracy = correct_num / 3000;