Out-of-Bag Error Estimation

Out-of-Bag Error

Built-in validation using samples left out of each tree

Training Set (n=100)

Select Tree to View

In-Bag samples 0

OOB samples 0

Each tree predicts its OOB samples

↓

Aggregate all OOB predictions

↓

Compare with true labels

↓

OOB Error: 0%

OOB error approximates test error without needing a separate validation set. ~37% of samples are OOB for each tree.