MatLab2012b/MatLab2013b分类器大全(svm,knn,随机森林等)

以你之姓@ 2022-06-17 00:25 233阅读 0赞

转载自：[http://blog.csdn.net/xuhaijiao99/article/details/15027093][http_blog.csdn.net_xuhaijiao99_article_details_15027093]

rain\_data是训练特征数据, train\_label是分类标签。  
Predict\_label是预测的标签。  
MatLab训练数据, 得到语义标签向量 Scores(概率输出)。  
  
1.逻辑回归(多项式MultiNomial logistic Regression)  
Factor = mnrfit(train\_data, train\_label);  
Scores = mnrval(Factor, test\_data);  
scores是语义向量(概率输出)。对高维特征，吃不消。  
  
2.随机森林分类器（Random Forest）  
Factor = TreeBagger(nTree, train\_data, train\_label);  
\[Predict\_label,Scores\] = predict(Factor, test\_data);  
scores是语义向量(概率输出)。实验中nTree = 500。  
效果好，但是有点慢。2500行数据，耗时400秒。500万行大数据分析，会咋样？准备好一篇小说慢慢阅读吧^\_^  
  
3.朴素贝叶斯分类（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. 支持向量机SVM分类  
Factor = svmtrain(train\_data, train\_label);  
predict\_label = svmclassify(Factor, test\_data);  
不能有语义向量 Scores(概率输出)  
  
支持向量机SVM(Libsvm)  
Factor = svmtrain(train\_label, train\_data, '-b 1');  
\[predicted\_label, accuracy, Scores\] = svmpredict(test\_label, test\_data, Factor, '-b 1');  
  
5.K近邻分类器 （KNN）  
predict\_label = knnclassify(test\_data, train\_data,train\_label, num\_neighbors);  
accuracy = length(find(predict\_label == test\_label))/length(test\_label)\*100;  
不能有语义向量 Scores(概率输出)  
  
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是近邻样本的下标集合，Dist是距离集合。  
自己编写, 实现概率输出 Scores(概率输出)  
  
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.集成学习器（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);  
效果比预期差了很多。不佳。  
  
7. 判别分析分类器（discriminant analysis classifier）  
Factor = ClassificationDiscriminant.fit(train\_data, train\_label);  
Factor = ClassificationDiscriminant.fit(train\_data, train\_label, 'discrimType', '判别类型:伪线性...');  
predict\_label = predict(Factor, test\_data);  
\[predict\_label, Scores\] = predict(Factor, test\_data);

[http_blog.csdn.net_xuhaijiao99_article_details_15027093]: http://blog.csdn.net/xuhaijiao99/article/details/15027093