AdaBoost : bootstrapping stumps

There are three main ideas behind the AdaBoost algorithm (in the case of trees)

1. AdaBoost combines a lot of "weak learners" to make classification. These weak learners are almost always "stumps" (i.e. trees with one root node and two leaves).
2. Some stumps get more say in the classification than others
3. Each stump is made by taking the previous stump's mistakes into account (via bootstrapping the sample following a distribution that puts more weight on incorrectly classified data)