R 中的蒙特卡洛交叉验证编码问题 - Auto MPG 数据集

问：

我只是在使用 Auto MPG 数据集在 R 中练习蒙特卡洛交叉验证编码问题。

在练习分类工具后，我只想应用交叉验证，设置 B = 100。 但是，我只是被 R 代码中的一些编码错误阻止了，如下所示;

均值误差（predict（mod1.简历，Auto_test。CV） ！= Auto_test。CV$mpg01） ： 不能强制“list”对象键入“integer”

警告：较长的物体长度不是较短物体长度的倍数警告：较长的物体长度不是较短物体长度的倍数[1]

请让我知道在哪里可以正确更改代码。谢谢！

full = rbind(Auto_train, Auto_test)
n1 = 274
set.seed(123)
n = dim(full[1])
B = 100
TEALL = NULL

for (b in 1:B){
  flag <- sort(sample(1:n,n1))
  Auto_train.CV <- full[flag,]
  Auto_test.CV <- full[-flag,]
  CV.test.err <- NULL
  
  mod1.CV <- lda(mpg01 ~ cylinders + displacement + horsepower + weight, data = Auto_train.CV)
  pred1.CV <- predict(mod1.CV, Auto_train.CV)$class
  CV.test.err <- cbind(CV.test.err, mean(predict(mod1.CV, Auto_test.CV) != Auto_test.CV$mpg01))
  
  mod2.CV <- qda(mpg01 ~ cylinders + displacement + horsepower + weight, data = Auto_train.CV)
  pred2.CV <- predict(mod2.CV, Auto_train.CV)$class
  CV.test.err <- cbind(CV.test.err, mean(predict(mod2.CV, Auto_test.CV) != Auto_test.CV$mpg01))
  
  mod3.CV <- naiveBayes(mpg01 ~ cylinders + displacement + horsepower + weight, data = Auto_train.CV)
  pred3.CV <- predict(mod3.CV, Auto_train.CV)
  CV.test.err <- cbind(CV.test.err, mean(predict(mod3.CV, Auto_test.CV) != Auto_test.CV$mpg01))

  mod4.CV <- multinom(mpg01 ~ cylinders + displacement + horsepower + weight, data = Auto_train.CV)
  pred4.CV <- predict(mod4.CV, Auto_train.CV)
  CV.test.err <- cbind(CV.test.err, mean(predict(mod4.CV, Auto_test.CV) != Auto_test.CV$mpg01))
 
  train.X.CV <- cbind(Auto_train.CV$cylinders, Auto_train.CV$displacement, Auto_train.CV$horsepower, Auto_train.CV$weight)
  test.X.CV <- cbind(Auto_test.CV$cylinders, Auto_test.CV$displacement, Auto_test.CV$horsepower, Auto_test.CV$weight)
  mod5.CV <- knn(train.X.CV, test.X.CV, Auto_train.CV$mpg01, k = 4)
  CV.test.err <- cbind(CV.test.err, mean(mod5.CV != Auto_test.CV$mpg01))
  
}

TEALL = rbind(TEALL, CV.test.err))