Difference between Bagging & Boosting and how these help




I recently came across the ensemble methods and there are two important terms, people are using more often, Bagging and Boosting.

Bagging: It is the method to decrease the variance of model by generating additional data for training from your original data set using combinations with repetitions to produce multisets of the same size as your original data.

Boosting: It helps to calculate the predict the target variables using different models and then average the result( may be using a weighted average approach).

Can you please help me with the difference between these two and when should we use which method? I came across multiple articles and most of these are saying that ensemble methods improves the power of model effectively. I would appreciate you for sharing the effect of these methods using an example.



i had the same question. This post has some pretty good explanation for this.