用R写一个stacking代码，先划分训练集和测试集，基模型用RF、SVM、XGBoost，对训练集做5折交叉验证，模型精度评价指标用R2、RMSE和MAE

以下是一个简单的stacking代码实现：

library(randomForest)
library(e1071)
library(xgboost)

# 读取数据
data <- read.csv("path/to/data.csv")

# 划分训练集和测试集
set.seed(123)
train_index <- sample(1:nrow(data), 0.7*nrow(data))
train_data <- data[train_index, ]
test_data <- data[-train_index, ]

# 定义基模型
rf_model <- randomForest(y ~ ., data = train_data)
svm_model <- svm(y ~ ., data = train_data)
xgb_model <- xgboost(data = as.matrix(train_data[, -1]), label = train_data$y, nrounds = 50, verbose = FALSE)

# 定义stacking模型
stacking_model <- lm(y ~ ., data = data.frame(matrix(NA, nrow = nrow(train_data), ncol = 3)))

# 5折交叉验证
folds <- createFolds(train_data$y, k = 5)
for (i in 1:5) {
  # 训练集和验证集
  train <- train_data[-folds[[i]], ]
  valid <- train_data[folds[[i]], ]
  
  # 预测结果作为新特征
  train$rf_pred <- predict(rf_model, train)
  valid$rf_pred <- predict(rf_model, valid)
  train$svm_pred <- predict(svm_model, train)
  valid$svm_pred <- predict(svm_model, valid)
  train$xgb