{8} revision 0 modified: 0-0-2006 0:0

Boost learning or AdaBoost - the idea is to update the discrete distribution used in training any algorithm to emphasize those points that are misclassified in the previous fit of a classifier. sensitive to outliers, but not overfitting.