2019-01-07[Stay Sharp]Bagging

what is Bagging?

bagging was propsed by Leo Breiman in 1994, also called Bootstrap aggregating, is a machine learning ensemble meta-algorithm designed to improve the stability and accuracy of machine learning. It help to avoid overfitting by reducing variance. It's usually applied to decision tree methods.

how does bagging work?

There are also two steps of bagging working process:

• generating training sets. generate new training sets by sampling from orginal training set uniformly and with replacement. If each new trainig size is equal to the original training set size and if is large, the has the fraction of the unique examples of .
  bagging.png
• training. train each new training set, generating the corresponding model. then use voting for classification or averaging for regression to select the model.
  bagging0.png

References

https://medium.com/open-machine-learning-course/open-machine-learning-course-topic-5-ensembles-of-algorithms-and-random-forest-8e05246cbba7

https://en.wikipedia.org/wiki/Bootstrap_aggregating

https://medium.com/open-machine-learning-course/open-machine-learning-course-topic-5-ensembles-of-algorithms-and-random-forest-8e05246cbba7