对于已经调好参数的模型，我们需要将其保存下来，可以使用下面两种方法。
模型训练
from sklearn.tree import DecisionTreeClassifier
from sklearn import datasets
X, y = datasets.load_iris(return_X_y=True)
model = DecisionTreeClassifier()
model.fit(X, y)
sklearn.externals.joblib
import joblib
joblib.dump(model, 'DecisionTree.pkl')
model = joblib.load('DecisionTree.pkl')
pickle
import pickle
f = open('DecisionTree.pickle', 'wb')
pickle.dump(model, f)
f.close()
f = open('DecisionTree.pickle', 'rb')
model = pickle.load(f)
f.close()
joblib与pickle对比
对于大数据而言，joblib比pickle更加高效，但是joblib只能将对象存储在磁盘文件中，不能保存为字符串。
cs
