Random Forest — Ensemble & OOB Error
Bootstrap aggregation of decision trees with out-of-bag error estimation
Controls
Trees in forest:
10
Max depth:
3
Points:
60
Grow Forest
New Dataset
Bagging:
Each tree trains on ~63% of data (bootstrap sample).
OOB Error:
Remaining ~37% are "out-of-bag" — used for unbiased error estimation without a validation set.
Trees grown:
0
OOB Accuracy:
—
Single tree acc:
—