Boosting and Bagging
Question
What is the difference between Boosting and Bagging?
Explai in detail
in progress
0
Machine Learning
55 years
1 Answer
822 views
Master 0
Answer ( 1 )
Bagging stands for Bootstrap aggregating. Bootstraping means creating multiple
datasets from a single dataset by performing sampling with replacement.
Generally, decision trees are associated with high variance. So, we build multiple
decision trees over bootstrapped samples and average out the results which help in reducing the
variance.
In boosting, we start with building weak learners and then go on combining the trees to
make one final strong learner. The observations which are misclassified in a particular
tree, are given more weightage while building the next tree. As we go on building more
and more trees, it leads to reduction in the error.