AdaBoost is a boosting algorithm based on the Gradient Descent optimization method. It fits a sequence of weak learners (originally stumps, that is, single-level decision trees) on re-weighted versions of the data. Weights are assigned based on the predictability of the case. Cases that are more difficult are weighted more. The idea is that the trees first learn easy examples and then concentrate more on the difficult ones. In the end, the sequence of weak learners is weighted to maximize the overall performance:

In: import numpy as np
    from sklearn.ensemble import AdaBoostClassifier
    hypothesis = AdaBoostClassifier(n_estimators=300, random_state=101)
    scores = cross_val_score(hypothesis, covertype_X, covertype_y, cv=3, 
                             scoring='accuracy', n_jobs=-1)
    print ("Adaboost -> cross validation accuracy: mean = %0.3f 
    std = %0.3f" % (np.mean(scores), np.std(scores)))
    Out: Adaboost -> cross validation accuracy: mean = 0.610 std = 0.014