14、监督机器学习与神经网络深入解析

监督机器学习与神经网络深入解析

1. 监督机器学习中的提升树算法

1.1 提升树的绘图实现

在监督机器学习中，有关于提升树的绘图函数 plot_boosting_trees 。以下是该函数的代码：

def plot_boosting_trees(data, steps=10, mcws=[30,20,20,10], gammas=
[100,200,300,500]):
    # to reduce clutter uncomment one of following two lines
    #mcws=[10]
    #gammas=[200]
    learners = [(mcw, gamma, Boosting_learner(data,
    sp_DT_learner(min_child_weight=mcw, gamma=gamma)))
    for gamma in gammas for mcw in mcws
    ]
    plt.ion()
    plt.xscale('linear') # change between log and linear scale
    plt.xlabel("number of trees")
    plt.ylabel("mean squared loss")
    markers = (m+c for c in ['k','g','r','b','m','c','y'] for m in
    ['-','--','-.',':'])
    for (mcw,gamma,learner) in