> pre1 <- predict(svm.model,kp_test,type='raw')
Error in match.arg(type, c("response", "probabilities", "votes", "decision")) :
'arg' should be one of “response”, “probabilities”, “votes”, “decision”
> pre1 <- predict(svm.model,kp_test,type='probabilities')
Error in .local(object, ...) :
ksvm object contains no probability model. Make sure you set the paramater prob.model in ksvm during training.
请教大家,我做了SVM模型,想画ROC曲线评估模型效果。
在树上看到naive bayes模型可以用type=“raw”,glm模型可以用type=“response”,但是svm模型的type应该用什么呢?这几种选择有何区别呢?
我尝试了svm也用type=“response”,但画出的ROC曲线和glm模型一模一样。这是什么道理,总感觉type不能随便选。
请教各位老师!!