Comparing bagging and boosting for natural language processing tasks, a typicality approach
Véronique Hoste, Walter M. P. Daelemans · 2000
We report on a comparison of two well-known ensemble combination methods, bagging and boosting, in the context of a natural language processing task. In comparative studies on general Machine Learning benchmark tasks, boosting has been shown to be more sensitive to noisy data than bagging. Because of the