如何在 R 中找到 K 折验证的平均训练和交叉验证错误？

Question

我正在运行一个决策树模型，我使用 K 折交叉验证将训练集随机分成 K 个不同的子集 (K = 10)

如何计算平均训练误差和平均交叉验证误差？

这是我目前的代码，我是机器学习的新手。

data_tree_copy <- data_tree[sample(nrow(data_tree)),]

## Create 10 equally size folds

k_folds <- cut(seq(1,nrow(data_tree_copy)),
             breaks = 10,
             labels = FALSE)

MSE_tree <- 0

for(i in 1:10){
  testIndexes <- which(k_folds == i, arr.ind=TRUE)
  testData <- data_tree_copy[testIndexes, ]
  trainData <- data_tree_copy[-testIndexes, ]
  
## median_value is my predictor.

  tree_model <- rpart(median_value~.,
                      data = trainData, 
                      method = "anova",
                      control = rpart.control(minsplit = 10, 
                                              cp = 0.001))

  predict2 <- predict(tree_model, testData)
  
  MSE_tree <- MSE_tree + sum(k_folds == i)/nrow(data_tree_copy) * mean((predict2 - testData$median_value)^2)

    }

RSME_pred  <- sqrt(MSE_tree)

print(paste("root mean-squared error of the prediction is", RSME_pred))


train_error <-  mean((predict2)^2)

train_error  

printcp(tree_model) # Cross-validation check

我在这里做错了什么！我得到了结果，但无法掌握。

提前致谢！！！