使用R的caret训练随机森林遇调参错误：需含mtry列

报错信息：

Error: The tuning parameter grid should have columns mtry

用户代码：

ctrlCV = trainControl(method = 'cv', number = 10 , classProbs = TRUE , savePredictions = TRUE, summaryFunction = twoClassSummary )

rfGRID = expand.grid(interaction.depth = c(2, 3, 5, 6, 7, 8, 10),
                     n.trees = c(50,75,100,125,150,200,250), 
                     shrinkage = seq(.005, .2,.005),
                     n.minobsinnode = c(5,7,10, 12 ,15, 20),
                     nodesize = c(1:10),
                     mtry = c(1:10))

RF_loop_trn = c()
RF_loop_tst = c()

for(i in (1:5)){
  print(i)
  
  IND = createDataPartition(y = scoresWithResponse$response, p=0.75, list = FALSE)
  scoresWithResponse.trn = scoresWithResponse[IND, ]
  scoresWithResponse.tst = scoresWithResponse[-IND,]
  
  rfFit = train(response~., data = scoresWithResponse.trn,
                importance = TRUE,
                method = "rf",
                metric="ROC",
                trControl = ctrlCV,
                tuneGrid = rfGRID,
                classProbs = TRUE,
                summaryFunction = twoClassSummary
  )
  
  
  RF_loop_trn[i] = auc(roc(scoresWithResponse.trn$response,predict(rfFit,scoresWithResponse.trn, type='prob')[,1]))
  RF_loop_tst[i] = ahaveroc(scoresWithResponse.tst$response,predict(rfFit,scoresWithResponse.tst, type='prob')[,1]))
  
}

用户已尝试的方案：从GitHub重新下载caret包、在expand.grid的参数前加.、仅给mtry加.前缀（如.mtry）、将mtry移至train函数中，但均出现相同错误。

错误原因 #

你构建的调参网格rfGRID混入了梯度提升树（gbm）的专属参数，比如interaction.depth、n.trees、shrinkage、n.minobsinnode，这些参数和随机森林（rf）完全不兼容。Caret中method="rf"对应的调参参数只有mtry，其他参数（如nodesize）属于模型的固定参数，不能放在tuneGrid里作为调参列。

当Caret解析你的调参网格时，大量无关参数干扰了它对当前模型（rf）对应调参列的识别，最终报错提示找不到mtry列。

解决步骤 #

1. 重构正确的调参网格
   只保留随机森林对应的调参参数mtry，示例：

   rfGRID = expand.grid(mtry = c(1:10))
   

   如果需要设置nodesize等固定参数，直接在train函数中通过额外参数传入，不需要放在tuneGrid里：

   rfFit = train(response~., data = scoresWithResponse.trn,
                 importance = TRUE,
                 method = "rf",
                 metric="ROC",
                 trControl = ctrlCV,
                 tuneGrid = rfGRID,
                 classProbs = TRUE,
                 summaryFunction = twoClassSummary,
                 nodesize = 5  # 这里设置固定的nodesize值
   )
   

2. 修正测试集AUC计算的笔误
   代码中ahaveroc是拼写错误，应该和训练集一样用auc(roc(...))：

   RF_loop_tst[i] = auc(roc(scoresWithResponse.tst$response,predict(rfFit,scoresWithResponse.tst, type='prob')[,1]))
   

3. 可选：扩展固定参数设置
   随机森林的其他参数（如树的数量ntree）同样作为固定参数传入train函数即可：

   rfFit = train(response~., data = scoresWithResponse.trn,
                 importance = TRUE,
                 method = "rf",
                 metric="ROC",
                 trControl = ctrlCV,
                 tuneGrid = rfGRID,
                 classProbs = TRUE,
                 summaryFunction = twoClassSummary,
                 ntree = 200,  # 固定树的数量
                 nodesize = 5
   )
   

内容的提问来源于stack exchange，提问作者Programming Noob