从随机森林中提取一棵树，然后使用提取的树进行预测

Question

As an example, let's use the iris data set.

library(randomForest)
data(iris)
smp_size <- floor(0.75 * nrow(iris))
train_ind <- sample(seq_len(nrow(iris)), size = smp_size)

train <- iris[train_ind, ]
test <- iris[-train_ind, ]

model <- randomForest(Species~., data = train, ntree=10)

If I use the getTree() function from the randomForest package, I can extract, for example, the third tree without any problem.

treefit <- getTree(model, 3)

But how can I use that (ie treefit) to make predictions on the test set, for instance? like "predict()", is there a function out there to do that directly?

Thank you in advance

Answer 1

You can use the predict function in the randomForest package directly by setting the predict.all argument to TRUE .

See the following reproducible code for how to use this: also see the help page for predict.randomForest here .

library(randomForest)
set.seed(1212)
x <- rnorm(100)
y <- rnorm(100, x, 10)
df_train <- data.frame(x=x, y=y)
x_test <- rnorm(20)
y_test <- rnorm(20, x_test, 10)
df_test <- data.frame(x = x_test, y = y_test)
rf_fit <- randomForest(y ~ x, data = df_train, ntree = 500)
# You get a list with the overall predictions and individual tree predictions
rf_pred <- predict(rf_fit, df_test, predict.all = TRUE)
rf_pred$individual[, 3] # Obtains the 3rd tree's predictions on the test data