关于使用R的随机森林的不同结果

Question

我拟合了随机森林模型，其中使用插入符号包进行交叉验证，发现每个分割处的变量数mtry被发现。 我的代码如下，

require(ISLR)
require(caret)
set.seed(123)
fitControl <- trainControl(method = "cv",number = 5,savePredictions = T)
mod_fit <- train(Direction ~ Lag1 + Lag2 + Lag3 + Lag4 + Lag5 + Volume,
                 data=Smarket, method = "rf", 
                 trControl = fitControl,tuneLength  =5,ntree=1000)



 > mod_fit$bestTune
  mtry
1    2
> 

我使用所有数据获得了最终模型，并且mtry = 2，如下所示，

> mod_fit$finalModel

Call:
 randomForest(x = x, y = y, ntree = 1000, mtry = param$mtry) 
               Type of random forest: classification
                     Number of trees: 1000
No. of variables tried at each split: 2

        OOB estimate of  error rate: 50.48%
Confusion matrix:
     Down  Up class.error
Down  265 337   0.5598007
Up    294 354   0.4537037

但是当我使用randomforest软件包对整个数据使用mtry = 2的相同模型时，我得到了不同的结果

require(randomForest)
set.seed(123)
rf <- randomForest(Direction ~ Lag1 + Lag2 + Lag3 + Lag4 + Lag5 + 
                     Volume, data = Smarket,  
                           mtry = 2, ntree = 1000, importance = TRUE)


    > rf

    Call:
     randomForest(formula = Direction ~ Lag1 + Lag2 + 
Lag3 + Lag4 +      Lag5 + Volume,
 data = Smarket, mtry = 2, ntree = 1000, importance = TRUE) 
                   Type of random forest: classification
                         Number of trees: 1000
    No. of variables tried at each split: 2

            OOB estimate of  error rate: 49.84%
    Confusion matrix:
         Down  Up class.error
    Down  270 332   0.5514950
    Up    291 357   0.4490741

结果不同的原因可能是什么？

同样，当我习惯预测功能时，我得到以下结果，即分类没有错误。

> rfPredict1 <- predict(mod_fit,newdata = Smarket )
> caret::confusionMatrix(rfPredict1, Smarket$Direction  )
Confusion Matrix and Statistics

          Reference
Prediction Down  Up
      Down  602   0
      Up      0 648

               Accuracy : 1          
                 95% CI : (0.9971, 1)
    No Information Rate : 0.5184     
    P-Value [Acc > NIR] : < 2.2e-16  

                  Kappa : 1          

 Mcnemar's Test P-Value : NA         

            Sensitivity : 1.0000     
            Specificity : 1.0000     
         Pos Pred Value : 1.0000     
         Neg Pred Value : 1.0000     
             Prevalence : 0.4816     
         Detection Rate : 0.4816     
   Detection Prevalence : 0.4816     
      Balanced Accuracy : 1.0000     

       'Positive' Class : Down  

我在这里做错了什么？ 谢谢