其核心思想是针对同一个训练集训练不同的分类器(弱分类器)，然后把这些弱分类器集合起来，构成一个更强的最终分类器(强分类器)。其算法本身是通过改变数据分布来实现的(Adaboost as an iterative algorithm which core idea is to train several different weak classifiers for the same dataset and then collect them together to form a stronger