Watch how Boosting sequentially combines simple "Weak Learners" (decision stumps) to build a powerful "Ensemble" classifier. Notice how the point sizes (sample weights) grow when they are misclassified, forcing the next learner to focus on them!
Shows a single split. Point size corresponds to current sample weight. Red borders indicate misclassified points.
Shows combined decision boundary of all stumps. The color intensity represents the confidence margin of the ensemble.