模型不支持libsvm中的概率估计 [英] Model does not support probabiliy estimates in libsvm
问题描述
在训练和预测过程中，我在Matlab中将libsvm与选项'-b 1'一起使用.但是它总是返回Model does not support probabiliy estimates，所以我没有得到任何概率或准确性估计.我在二进制类SVM(不是nu-svm！)中尝试过，它应该与'-b 1'一起工作，但事实并非如此.有谁知道这个问题的原因是什么?
I used libsvm in Matlab with the option '-b 1' in both training and prediction process. But it always returns Model does not support probabiliy estimates, so I don't get any probability or accuracy estimation. I tried in binary class SVM (not nu-svm!), it should have work with the '-b 1' but it's not. Does anyone know what's the reason for this problem?
谢谢
推荐答案
- 让我告诉您 svm-predict 的用法:
- Let me show you the usage of svm-predict:
用法:svm-predict [选项] test_file model_file output_file 选项:
Usage: svm-predict [options] test_file model_file output_file options:
-b potential_estimates:是预测概率估计值，是0还是1(默认值为0)；对于一类SVM，仅支持0
-b probability_estimates: whether to predict probability estimates, 0 or 1 (default 0); for one-class SVM only 0 is supported
-q:安静模式(无输出)
-q : quiet mode (no outputs)
- svm-train:
用法:svm-train [选项] training_set_file [model_file]选项:
Usage: svm-train [options] training_set_file [model_file] options:
-s svm_type : set type of SVM (default 0)
0-C-SVC(多类别分类)
0 -- C-SVC (multi-class classification)
1-nu-SVC(多类别分类)
1 -- nu-SVC (multi-class classification)
2-一类SVM
3-epsilon-SVR(回归)4-nu-SVR(回归)
3 -- epsilon-SVR (regression) 4 -- nu-SVR (regression)
-t kernel_type:设置内核函数的类型(默认为2)
0-线性:u' v 1-多项式:(gamma u' v + coef0)^ degree
2-径向基函数:exp(-gamma | u-v | ^ 2)
3-乙状结肠:tanh(gamma * u'* v + coef0)
-t kernel_type : set type of kernel function (default 2)
0 -- linear: u'v 1 -- polynomial: (gammau'v + coef0)^degree
2 -- radial basis function: exp(-gamma|u-v|^2)
3 -- sigmoid: tanh(gamma*u'*v + coef0)
4-预先计算的内核(training_set_file中的内核值)
4 -- precomputed kernel (kernel values in training_set_file)
-d度:设置内核函数的度(默认为3)
-d degree : set degree in kernel function (default 3)
-g gamma:在内核函数中设置gamma(默认值为1/num_features)
-g gamma : set gamma in kernel function (default 1/num_features)
-r coef0:在内核函数中设置coef0(默认为0)
-r coef0 : set coef0 in kernel function (default 0)
-c cost:设置C-SVC，epsilon-SVR和nu-SVR的参数C(默认为1)
-c cost : set the parameter C of C-SVC, epsilon-SVR, and nu-SVR (default 1)
-n nu:设置nu-SVC，一类SVM和nu-SVR的参数nu(默认值为0.5)
-n nu : set the parameter nu of nu-SVC, one-class SVM, and nu-SVR (default 0.5)
-p epsilon:在epsilon-SVR的损失功能中设置epsilon(默认值为0.1)
-p epsilon : set the epsilon in loss function of epsilon-SVR (default 0.1)
-m cachesize:以MB为单位设置缓存内存大小(默认为100)
-m cachesize : set cache memory size in MB (default 100)
-e epsilon:设置终止标准的容差(默认为0.001)
-e epsilon : set tolerance of termination criterion (default 0.001)
-h收缩:是使用0还是1(默认为1)的收缩试探法
-h shrinking : whether to use the shrinking heuristics, 0 or 1 (default 1)
-b density_estimates:是否为概率估计训练SVC或SVR模型，即0还是1(默认为0)
-wi weight:对于C-SVC(默认值为1)，将类i的参数C设置为weight * C.
-wi weight : set the parameter C of class i to weight*C, for C-SVC (default 1)
-v n:n折交叉验证模式
-v n: n-fold cross validation mode
-q:安静模式(无输出)
-q : quiet mode (no outputs)
我们可以看到最后第四行是-b选项.如果我们使用"-b 1"选项训练模型，那么您将获得一个可以在您尝试进行预测时输出概率的模型.否则，如果您仅在尝试预测而不使用'-b 1'生成模型时使用'-b 1'选项.您会得到错误:模型不支持概率估计
We can see that the last fourth line is -b option. If we trained the model with '-b 1' option, we'll get a model that can output probability when you try to predict. Otherwise, if you only use '-b 1' option when you try to predict and not generate a model with '-b 1'. you will get the error : Model does not support probabiliy estimates
最主要的是，如果要获取概率估计值，则应在火车中使用" -b 1 "并进行测试过程，它们两个.
The main thing is that if you want to get probabiliy estimates, you should use '-b 1' in your train and test process, both of them.
这篇关于模型不支持libsvm中的概率估计的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！
