捕获随机森林

Question

我正在运行一个袋装的随机森林模型，我无法捕获每棵树中使用的预测变量。我总共有34个预测变量，我试图运行一个带有4个预测变量的树模型。对于生成的每个树模型，我想捕获预测变量并将它们保存到名为＆＃34; lstVariableUsed的项目中。＆＃34;

我遇到线路问题＆＃34; lstVariablesUsed＆lt; - rbind（lstVariablesUsed，varUsed（model.bagged））。＆＃34;

我已尝试过多种方法来捕获它，包括varUsed，model.bagged $ terms，Name（）和attr（），但是没有运气获得每棵树中使用的4个变量。任何帮助表示赞赏。

set.seed(1)

lstVariablesUsed <- character()

maxnumpred <- 4
maxnumtrees <- 14


for(numpreds in 1:maxnumpred){
  for(numtrees in 1:maxnumtrees){

    train <- sample(1:nrow(df.FinalClass), nrow(df.FinalClass)/2) 
    FinalClass.test2 = df.FinalClass[-train,]
    DidRetain.test2 = DidRetain[-train]

    model.bagged <- randomForest(DidRetain~., 
                             data=df.FinalClass, 
                             subset=train, 
                             mtry=maxnumpred,
                             ntree=maxnumtrees,
                             importance = FALSE)

    pred.vals.tree <- predict(model.bagged, FinalClass.test2, type="class")
    t <- table(pred.vals.tree, DidRetain.test2)

    acc.val <- (sum(diag(t))/sum(t))*100

    #store the accuracy value in the lstAccuraccy's
    lstAccuracy <- rbind(lstAccuracy, acc.val) 

    #store the variables used in the model
    lstVariablesUsed <- rbind(lstVariablesUsed, varUsed(model.bagged))

    #output the number of iterations of for loop that have passed
    print(paste("     Processed trees: ", numtrees))
  }
#output the number of iterations of for loop that have passed
print(paste("Processed predictors: ", numpreds))
}