男女性别预测

# [height, weight, shoe size]
X = [[181, 80, 44], [177, 70, 43], [160, 60, 38], [154, 54, 37],
     [166, 65, 40], [190, 90, 47], [175, 64, 39], [177, 70, 40],
     [159, 55, 38], [171, 75, 42], [181, 85, 43], [148, 70, 42]]

y = ['male', 'female', 'female', 'female',
         'female', 'female', 'female', 'female',
         'male', 'female', 'male', 'male']

"""选择决策树算法，训练算法"""
from sklearn import tree
try:
    with open('DecisionTree.pickle', 'rb') as f:
        clf = pickle.load(f)
except Exception, e:
    # 训练算法
    clf = tree.DecisionTreeClassifier()
    clf.fit(X, y)

    # 序列化算法
    with open('DecisionTree.pickle', 'wb') as f:
        import pickle
        pickle.dump(clf, f)

# make a prediction.
prediction = clf.predict([[190, 70, 43], [156, 60, 36]])
print prediction
```    


```python
"""Visualization: graphviz export of the above tree trained on the entire dataset;
the results are saved in an output file GenderClassifier.pdf"""

# Graphviz是图形绘制工具,可以很方便的用来绘制结构化的图形网络,支持多种格式输出
import graphviz 
dot_data = tree.export_graphviz(clf, out_file=None,
                                feature_names=['height', 'weight', 'shoe size'],
                                class_names='gender',
                                filled=True, rounded=True,
                                special_characters=True)
graph = graphviz.Source(dot_data)
graph.render('GenderClassifier')