我正在尝试使用 libsvm（通过 Matlab 接口）来运行一些多标签分类问题。这里有一个使用 IRIS 数据的示例问题：

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 classifieroptsStruct                              = ['-c ', num2str(2), ' -g ', num2str(4), '-b ', 1];SVMClassifierObject                     = svmtrain(groundTruthGroupNumTraining, featureSelectedTraining, optsStruct);optsStruct                              = ['-b ', 1];[predLabelTesting, predictAccuracyTesting, ...    predictScoresTesting]               = svmpredict(groundTruthGroupNumTesting, featureSelectedTesting, SVMClassifierObject, optsStruct);

然而，对于我得到的预测概率（这里展示了前12行的结果）

1.08812899093155    1.09025554950852    -0.01400090569120010.948911671379753   0.947899227815959   -0.01400090569260240.521486301840914   0.509673405799383   -0.01400090569260270.914684487894784   0.912534150299246   -0.01400090569260271.17426551505833    1.17855350325579    -0.01400090569251030.567801459258613   0.557077025701113   -0.01400090569260270.506405203427106   0.494342606399178   -0.01400090569260270.930191457490471   0.928343421250020   -0.01400090569260271.16990617214906    1.17412523596840    -0.01400090569260261.16558843984163    1.16986137054312    -0.01400090569260150.879648874624610   0.876614924593740   -0.0140009056926027-0.151223818963057  -0.179682730685229  -0.0140009056925999

我感到困惑的是，为什么有些概率大于1，有些是负数？

然而，预测的标签似乎相当准确：

111111111113

最终输出为

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

那么如何解释预测概率的结果呢？非常感谢。A.

回答：

SVM 的输出不是概率！

分数的符号表示它属于类别 A 还是类别 B。如果分数是 1 或 -1，它就在边界上，尽管知道这一点并不是特别有用。

如果你真的需要概率，你可以使用 Platt 缩放 将它们转换。你基本上是对它们应用一个 sigmoid 函数。