August 2019
Intermediate to advanced
342 pages
9h 35m
English
The term bootstrap refers to the operation of sampling with a replacement that's been applied to a dataset. The bagging method, therefore, associates an individual estimator with each bootstrap; the ensemble estimator is implemented by applying the majority voting method to individual classifiers.
The number of bootstraps to be taken into consideration can be predetermined or adjusted using a validation dataset.
The bagging method is particularly useful in the case where sampling with replacement helps to rebalance the original dataset, thus reducing total variance.
Read now
Unlock full access