0.25
0.5
0.75
1.25
1.5
1.75
2
AdaBoost is Universally Consistent
Published on Feb 25, 20078243 Views
We consider the risk, or probability of error, of the classifier produced by AdaBoost, and in particular the stopping strategy to be used to ensure universal consistency. (A classification method is u
Related categories
Chapter list
AdaBoost is Universally Consistent00:00
AdaBoost00:30
Universal Consistency01:55
AdaBoost is Universally Consistent03:26
Previous results: Regularized versions04:24
Previous results: Regularized versions 0106:06
Previous results: Bounded step size08:05
Previous results about AdaBoost09:05
Previous result about AdaBoost: ‘Process consistency’10:05
AdaBoost is Universally Consistent12:19
The key theorem12:26
The key theorem 0114:21
The key theorem: Idea of proof15:55
The key theorem: Idea of proof 0119:33
The key theorem: Idea of proof 0220:05
The key theorem: Idea of proof 0320:30
The key theorem: Idea of proof 0421:30
Open Problems22:36
AdaBoost is Universally Consistent25:01