Deep Learning
Bagging is the concept of splitting a dataset and randomly placing it into bags for training a model.
False
Proposed by @svpino
Card 1 of 259. Answer: False
False
Proposed by @svpino
Card 1 of 259. Answer: False
About · Quizzes · History of AI · To suggest new questions or report an error send me a dm.
False, as worded. Bagging, short for bootstrap aggregating, does not split the dataset into separate bags. It draws many bootstrap samples, each as large as the original data and drawn with replacement, trains one model per sample, and combines their predictions by averaging or majority vote. Leo Breiman published it in 1996.
Because sampling is with replacement, the bags overlap and repeat rows. The chance that a given row is never picked in n draws is (1 − 1/n)ⁿ, which approaches 1/e ≈ 0.368, so each bag holds about 63.2% of the distinct rows; the left-out, out-of-bag rows give a free validation estimate. Averaging models trained on slightly different data cancels much of their variance.
Bagging helps most with unstable, high-variance learners such as deep decision trees; a random forest is bagged trees plus a random subset of features at each split. Interviewers often contrast it with boosting, which trains models one after another on the previous ones' errors and mainly reduces bias. Sampling without replacement instead is called pasting.