A gradient boosting model with 1000 trees and learning rate 0.01 shows overfitting. Which regularization approach is most effective without reducing tree count?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because stochastic gradient boosting (row subsampling) introduces randomness by training each tree on a random subset of observations, which acts as regularization and reduces variance. A (lower learning rate) slows learning but requires more trees for equal performance.
Full explanation below image
Full Explanation
B is correct because stochastic gradient boosting (row subsampling) introduces randomness by training each tree on a random subset of observations, which acts as regularization and reduces variance. A (lower learning rate) slows learning but requires more trees for equal performance. C (more depth) increases complexity and worsens overfitting. D (more trees at same rate) can increase overfitting.