R语言决策树AUC交叉验证出现Y0[[v]]索引越界报错如何解决

核心报错诱因 #

你遇到的索引越界问题主要来自三处错误：

• .doFit函数参数大小写定义错误：函数第一个参数定义为大写V，但内部同时用了大写V和小写v取值，调用时传入的是循环变量小写v，变量匹配混乱
• 子函数未传入y参数：主函数的y变量不会自动传入并行环境下的.doFit函数，导致列索引计算错误
• 缺少边界校验：如果数据集中某一类（0/1）的样本数量小于交叉验证折数V，split生成的类折列表长度会小于V，直接取值就会触发越界

修复代码 #

1 修复.cvFolds函数，增加边界校验 #

.cvFolds <- function(Y, V) {  
  # 提前校验类样本量，避免拆分后折数不足
  n_zero <- sum(Y == 0)
  n_one <- sum(Y == 1)
  if (n_zero < V | n_one < V) {
    stop("0类或1类样本数量不能小于交叉验证折数V，请减小V或扩充数据集")
  }
  Y0 <- split(sample(which(Y == 0)), rep(1:V, length = n_zero))
  Y1 <- split(sample(which(Y == 1)), rep(1:V, length = n_one))
  folds <- vector("list", length = V)
  for (v in seq(V)) folds[[v]] <- c(Y0[[v]], Y1[[v]])   
  return(folds)
}

2 修复.doFit函数的参数和变量引用 #

# 统一参数为小写v，新增y参数传入
.doFit <- function(v, folds, train, y) {
  set.seed(v)
  ycol <- which(names(train) == y)
  params <- list(x     = train[-folds[[v]], -ycol],
                 y     = as.factor(train[-folds[[v]], ycol]),
                 xtest = train[folds[[v]], -ycol])
  fit <- do.call(randomForest, params)
  pred <- fit$test$votes[, 2]
  return(pred)
}

3 修复主函数的子函数调用逻辑 #

iid_example <- function(train, y = "V1", V = 10, seed = 1) {
  set.seed(seed)
  folds <- .cvFolds(Y = train[, c(y)], V = V)
  
  # Generate CV predicted values
  cl <- makeCluster(detectCores())
  registerDoParallel(cl)
  # 调用时传入y参数
  predictions <- foreach(v = 1:V, .combine = "c", 
    .packages = c("randomForest")) %dopar% .doFit(v, folds, train, y)
  stopCluster(cl)
  predictions[unlist(folds)] <- predictions

  # Get CV AUC
  runtime <- system.time(res <- ci.cvAUC(predictions = predictions, 
                                         labels = train[, c(y)],
                                         folds = folds, 
                                         confidence = 0.95))
  print(runtime)
  return(res)
}

使用说明 #

如果你的数据集正负样本数小于10，可以将调用时的V参数调整为更小的数值，比如5折交叉验证：

res <- iid_example(train = datos, y = "V1", V = 5, seed = 1)

内容的提问来源于stack exchange，提问作者Jesús Asdrúbal Molina Vízquez