Question

Random Forest typically achieves lower variance than a single, fully-grown decision tree primarily because of which mechanism?

Q: Answer-Random Forest typically achieves lower variance than a single, fully-grown decision tree primarily b...

Random Forest builds many decision trees, each on an independent bootstrap sample of the training data, and additionally restricts each split to consider only a random subset of features. This deliberate injection of randomness decorrelates the individual trees they make different kinds of errors, so that averaging their predictions cancels out much of each trees individual variance while largely preserving low bias unlike boosting option D, which builds trees sequentially and adjusts weights based on prior errors.

A It uses gradient descent to iteratively minimize training loss across all trees jointly

B It prunes each individual tree far more aggressively than a standalone decision tree would be

C It averages predictions from many trees, each trained on a bootstrap sample with a random feature subset at each split, reducing inter-tree correlation

D It sequentially reweights misclassified samples between successive trees, similar to boosting

E It replaces the Gini impurity criterion with cross-entropy loss for all trees

Practice Next