Bagging (e.g., Random Forest) vs boosting (e.g., GBM) differs in that:

Question

Accepted Answer

Bagging averages independent learners to reduce variance; boosting sequentially focuses on residuals/errors — Bagging // variance via averages; boosting sequences residuals. Different ensembles.

Answer

Bagging always uses neural nets

Answer

Boosting cannot overfit

Answer

They are mathematically identical always

Bagging (e.g., Random Forest) vs boosting (e.g., GBM) differs in that:

Short Explanation and Infographic

Full Explanation