MATLAB2013B 分類器大全

MATLAB圖像識別 2017-07-18

展開全文

train_data是訓練特征數(shù)據(jù), train_label是分類標簽。
Predict_label是預測的標簽。
MatLab訓練數(shù)據(jù), 得到語義標簽向量 Scores(概率輸出)。

1.邏輯回歸(多項式MultiNomial logistic Regression)
Factor = mnrfit(train_data, train_label);
Scores = mnrval(Factor, test_data);
scores是語義向量(概率輸出)。對高維特征，吃不消。

2.隨機森林分類器（Random Forest）
Factor = TreeBagger(nTree, train_data, train_label);
[Predict_label,Scores] = predict(Factor, test_data);
scores是語義向量(概率輸出)。實驗中nTree = 500。
效果好，但是有點慢。2500行數(shù)據(jù)，耗時400秒。500萬行大數(shù)據(jù)分析，會咋樣？準備好一篇小說慢慢閱讀吧^_^

3.樸素貝葉斯分類（Naive Bayes）
Factor = NaiveBayes.fit(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;
效果不佳。

4. 支持向量機SVM分類
Factor = svmtrain(train_data, train_label);
predict_label = svmclassify(Factor, test_data);
不能有語義向量 Scores(概率輸出)

支持向量機SVM(Libsvm)
Factor = svmtrain(train_label, train_data, '-b 1');
[predicted_label, accuracy, Scores] = svmpredict(test_label, test_data, Factor, '-b 1');

5.K近鄰分類器（KNN）
predict_label = knnclassify(test_data, train_data,train_label, num_neighbors);
accuracy = length(find(predict_label == test_label))/length(test_label)*100;
不能有語義向量 Scores(概率輸出)

IDX = knnsearch(train_data, test_data);
IDX = knnsearch(train_data, test_data, 'K', num_neighbors);
[IDX, Dist] = knnsearch(train_data, test_data, 'K', num_neighbors);
IDX是近鄰樣本的下標集合，Dist是距離集合。
自己編寫, 實現(xiàn)概率輸出 Scores(概率輸出)

Matlab 2012新版本:
Factor = ClassificationKNN.fit(train_data, train_label, 'NumNeighbors', num_neighbors);
predict_label = predict(Factor, test_data);
[predict_label, Scores] = predict(Factor, test_data);

6.集成學習器（Ensembles for Boosting, Bagging, or Random Subspace）
Matlab 2012新版本:
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 = predict(Factor, test_data);
[predict_label, Scores] = predict(Factor, test_data);
效果比預期差了很多。不佳。

7. 判別分析分類器（discriminant analysis classifier）
Factor = ClassificationDiscriminant.fit(train_data, train_label);
Factor = ClassificationDiscriminant.fit(train_data, train_label, 'discrimType', '判別類型:偽線性...');
predict_label = predict(Factor, test_data);

[predict_label, Scores] = predict(Factor, test_data);