import numpy as np

from matplotlib import  pyplot as plt
from sklearn import neighbors, datasets
from matplotlib.colors import ListedColormap
from sklearn.neural_network import MLPClassifier

## 加载数据集
np.random.seed(0)
# 使用 scikit-learn  自带的 iris 数据集
iris=datasets.load_iris() 
# 使用前两个特征，方便绘图
X=iris.data[:,0:2] 
# 标记值
Y=iris.target 
data=np.hstack((X,Y.reshape(Y.size,1)))
# 混洗数据。因为默认的iris 数据集：前50个数据是类别0，中间50个数据是类别1，末尾50个数据是类别2.混洗将打乱这个顺序
np.random.shuffle(data) 
X=data[:,:-1]
Y=data[:,-1]
train_x=X[:-30]
train_y=Y[:-30]
# 最后30个样本作为测试集
test_x=X[-30:] 
test_y=Y[-30:]
    
def plot_classifier_predict_meshgrid(ax,clf,x_min,x_max,y_min,y_max):
    '''
     绘制 MLPClassifier 的分类结果

    :param ax:  Axes 实例，用于绘图
    :param clf: MLPClassifier 实例
    :param x_min: 第一维特征的最小值
    :param x_max: 第一维特征的最大值
    :param y_min: 第二维特征的最小值
    :param y_max: 第二维特征的最大值
    :return: None
      '''
    plot_step = 0.02 # 步长
    xx, yy = np.meshgrid(np.arange(x_min, x_max, plot_step),np.arange(y_min, y_max, plot_step))
    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])
    Z = Z.reshape(xx.shape)
    # 绘图
    ax.contourf(xx, yy, Z, cmap=plt.cm.Paired) 
        
def plot_samples(ax,x,y):
    '''
        绘制二维数据集

        :param ax:  Axes 实例，用于绘图
        :param x: 第一维特征
        :param y: 第二维特征
        :return: None
      '''
    n_classes = 3
    # 颜色数组。每个类别的样本使用一种颜色
    plot_colors = "bry" 
    for i, color in zip(range(n_classes), plot_colors):
        idx = np.where(y == i)
        # 绘图
        ax.scatter(x[idx, 0], x[idx, 1], c=color,label=iris.target_names[i], cmap=plt.cm.Paired)
        
def mlpclassifier_iris():
    '''
    使用 MLPClassifier 预测调整后的 iris 数据集
    '''
    fig=plt.figure()
    ax=fig.add_subplot(1,1,1)
    classifier=MLPClassifier(activation='logistic',max_iter=10000,hidden_layer_sizes=(30,))
    classifier.fit(train_x,train_y)
    train_score=classifier.score(train_x,train_y)
    test_score=classifier.score(test_x,test_y)
    x_min, x_max = train_x[:, 0].min() - 1, train_x[:, 0].max() + 2
    y_min, y_max = train_x[:, 1].min() - 1, train_x[:, 1].max() + 2
    plot_classifier_predict_meshgrid(ax,classifier,x_min,x_max,y_min,y_max)
    plot_samples(ax,train_x,train_y)
    ax.legend(loc='best')
    ax.set_xlabel(iris.feature_names[0])
    ax.set_ylabel(iris.feature_names[1])
    ax.set_title("train score:%f;test score:%f"%(train_score,test_score))
    plt.show()
    
mlpclassifier_iris()

def mlpclassifier_iris_hidden_layer_sizes():
    '''
    使用 MLPClassifier 预测调整后的 iris 数据集。考察不同的 hidden_layer_sizes 的影响

    :return: None
    '''
    fig=plt.figure()
    # 候选的 hidden_layer_sizes 参数值组成的数组
    hidden_layer_sizes=[(10,),(30,),(100,),(5,5),(10,10),(30,30)] 
    for itx,size in enumerate(hidden_layer_sizes):
        ax=fig.add_subplot(2,3,itx+1)
        classifier=MLPClassifier(activation='logistic',max_iter=10000,hidden_layer_sizes=size)
        classifier.fit(train_x,train_y)
        train_score=classifier.score(train_x,train_y)
        test_score=classifier.score(test_x,test_y)
        x_min, x_max = train_x[:, 0].min() - 1, train_x[:, 0].max() + 2
        y_min, y_max = train_x[:, 1].min() - 1, train_x[:, 1].max() + 2
        plot_classifier_predict_meshgrid(ax,classifier,x_min,x_max,y_min,y_max)
        plot_samples(ax,train_x,train_y)
        ax.legend(loc='best')
        ax.set_xlabel(iris.feature_names[0])
        ax.set_ylabel(iris.feature_names[1])
        ax.set_title("layer_size:%s;train score:%f;test score:%f"%(size,train_score,test_score))
    plt.show()
    
mlpclassifier_iris_hidden_layer_sizes()

def mlpclassifier_iris_ativations():
    '''
    使用 MLPClassifier 预测调整后的 iris 数据集。考察不同的 activation 的影响
    '''
    fig=plt.figure()
    # 候选的激活函数字符串组成的列表
    ativations=["logistic","tanh","relu"] 
    for itx,act in enumerate(ativations):
        ax=fig.add_subplot(1,3,itx+1)
        classifier=MLPClassifier(activation=act,max_iter=10000,hidden_layer_sizes=(30,))
        classifier.fit(train_x,train_y)
        train_score=classifier.score(train_x,train_y)
        test_score=classifier.score(test_x,test_y)
        x_min, x_max = train_x[:, 0].min() - 1, train_x[:, 0].max() + 2
        y_min, y_max = train_x[:, 1].min() - 1, train_x[:, 1].max() + 2
        plot_classifier_predict_meshgrid(ax,classifier,x_min,x_max,y_min,y_max)
        plot_samples(ax,train_x,train_y)
        ax.legend(loc='best')
        ax.set_xlabel(iris.feature_names[0])
        ax.set_ylabel(iris.feature_names[1])
        ax.set_title("activation:%s;train score:%f;test score:%f"%(act,train_score,test_score))
    plt.show()
    
mlpclassifier_iris_ativations()

def mlpclassifier_iris_algorithms():
    '''
    使用 MLPClassifier 预测调整后的 iris 数据集。考察不同的 algorithm 的影响

    :return: None
    '''
    fig=plt.figure()
    algorithms=["lbfgs","sgd","adam"] # 候选的算法字符串组成的列表
    for itx,algo in enumerate(algorithms):
        ax=fig.add_subplot(1,3,itx+1)
        classifier=MLPClassifier(activation="tanh",max_iter=10000,hidden_layer_sizes=(30,),solver=algo)
        classifier.fit(train_x,train_y)
        train_score=classifier.score(train_x,train_y)
        test_score=classifier.score(test_x,test_y)
        x_min, x_max = train_x[:, 0].min() - 1, train_x[:, 0].max() + 2
        y_min, y_max = train_x[:, 1].min() - 1, train_x[:, 1].max() + 2
        plot_classifier_predict_meshgrid(ax,classifier,x_min,x_max,y_min,y_max)
        plot_samples(ax,train_x,train_y)
        ax.legend(loc='best')
        ax.set_xlabel(iris.feature_names[0])
        ax.set_ylabel(iris.feature_names[1])
        ax.set_title("algorithm:%s;train score:%f;test score:%f"%(algo,train_score,test_score))
    plt.show()
    
mlpclassifier_iris_algorithms()

def mlpclassifier_iris_eta():
    '''
    使用 MLPClassifier 预测调整后的 iris 数据集。考察不同的学习率的影响
    '''
    fig=plt.figure()
    etas=[0.1,0.01,0.001,0.0001] # 候选的学习率值组成的列表
    for itx,eta in enumerate(etas):
        ax=fig.add_subplot(2,2,itx+1)
        classifier=MLPClassifier(activation="tanh",max_iter=1000000,
        hidden_layer_sizes=(30,),solver='sgd',learning_rate_init=eta)
        classifier.fit(train_x,train_y)
        iter_num=classifier.n_iter_
        train_score=classifier.score(train_x,train_y)
        test_score=classifier.score(test_x,test_y)
        x_min, x_max = train_x[:, 0].min() - 1, train_x[:, 0].max() + 2
        y_min, y_max = train_x[:, 1].min() - 1, train_x[:, 1].max() + 2
        plot_classifier_predict_meshgrid(ax,classifier,x_min,x_max,y_min,y_max)
        plot_samples(ax,train_x,train_y)
        ax.legend(loc='best')
        ax.set_xlabel(iris.feature_names[0])
        ax.set_ylabel(iris.feature_names[1])
        ax.set_title("eta:%f;train score:%f;test score:%f;iter_num:%d"%(eta,train_score,test_score,iter_num))
    plt.show()
    
mlpclassifier_iris_eta()