RandomForest调用报错：'User ID'未找到，如何忽略该列训练模型？

The error happens because your formula Purchased~. tells randomForest to use every column in the Network dataset as a predictor—including User ID. Since User ID has a space in its name, R can't parse it correctly without special handling, and anyway, we don't want this unique identifier cluttering up our model. Here are three straightforward fixes:

1. Modify the Formula to Exclude User ID #

This is the simplest approach. Use the minus sign (-) to remove the User ID column from the predictor set, and wrap the column name in backticks because it contains a space:

library(randomForest)
rfModel2 <- randomForest(
  formula = Purchased ~ . - `User ID`,
  data = Network,
  ntree = 50,
  importance = TRUE,
  replace = TRUE
)

2. Subset the Dataset Before Training #

If you prefer working with a stripped-down version of your data, you can subset the dataframe directly in the data argument to drop User ID:

library(randomForest)
# Remove User ID by name
rfModel2 <- randomForest(
  formula = Purchased~.,
  data = Network[, !names(Network) %in% "User ID"],
  ntree = 50,
  importance = TRUE,
  replace = TRUE
)

3. Use Tidyverse Syntax (If You're Comfortable With It) #

If you use the dplyr package, you can use select() to cleanly drop the unwanted column:

library(randomForest)
library(dplyr)

rfModel2 <- randomForest(
  formula = Purchased~.,
  data = Network %>% select(-`User ID`),
  ntree = 50,
  importance = TRUE,
  replace = TRUE
)

All three methods will ensure your model only uses relevant predictors to train on the 1/0 Purchased target variable, while avoiding the missing object error from the spaced column name.

内容的提问来源于stack exchange，提问作者Denisse Kowaleski