matlab - 多标签分类的 libsvm 输出预测概率

Question

我正在尝试使用 libsvm (带有 Matlab 接口(interface))来运行一些多标签分类问题。这是使用 IRIS 数据的一些玩具问题:

load fisheriris;

featuresTraining                        = [meas(1:30,:); meas(51:80,:); meas(101:130,:)];
featureSelectedTraining                 = featuresTraining(:,1:3);

groundTruthGroupTraining                = [species(1:30,:); species(51:80,:); species(101:130,:)];
[~, ~, groundTruthGroupNumTraining]     = unique(groundTruthGroupTraining);

featuresTesting                         = [meas(31:50,:); meas(81:100,:); meas(131:150,:)];
featureSelectedTesting                  = featuresTesting(:,1:3);

groundTruthGroupTesting                 = [species(31:50,:); species(81:100,:); species(131:150,:)];
[~, ~, groundTruthGroupNumTesting]      = unique(groundTruthGroupTesting);

% Train the classifier
optsStruct                              = ['-c ', num2str(2), ' -g ', num2str(4), '-b ', 1];
SVMClassifierObject                     = svmtrain(groundTruthGroupNumTraining, featureSelectedTraining, optsStruct);

optsStruct                              = ['-b ', 1];
[predLabelTesting, predictAccuracyTesting, ...
    predictScoresTesting]               = svmpredict(groundTruthGroupNumTesting, featureSelectedTesting, SVMClassifierObject, optsStruct);

但是，对于我得到的预测概率(此处显示前 12 行结果)

1.08812899093155    1.09025554950852    -0.0140009056912001
0.948911671379753   0.947899227815959   -0.0140009056926024
0.521486301840914   0.509673405799383   -0.0140009056926027
0.914684487894784   0.912534150299246   -0.0140009056926027
1.17426551505833    1.17855350325579    -0.0140009056925103
0.567801459258613   0.557077025701113   -0.0140009056926027
0.506405203427106   0.494342606399178   -0.0140009056926027
0.930191457490471   0.928343421250020   -0.0140009056926027
1.16990617214906    1.17412523596840    -0.0140009056926026
1.16558843984163    1.16986137054312    -0.0140009056926015
0.879648874624610   0.876614924593740   -0.0140009056926027
-0.151223818963057  -0.179682730685229  -0.0140009056925999

我很困惑，为什么有些概率大于 1，有些概率却为负？

但是，预测的标签似乎相当准确:

1
1
1
1
1
1
1
1
1
1
1
3

最终输出

Accuracy = 93.3333% (56/60) (classification)

那么如何解释预测概率的结果呢？多谢。答:

Answer 1

支持向量机的输出不是概率!

分数的符号表示它属于 A 类还是 B 类。如果分数是 1 或 -1，则它处于边缘，尽管知道这一点并不是特别有用。

如果您确实需要概率，可以使用 Platt scaling 进行转换。 。您基本上对它们应用了 sigmoid 函数。