模型保存和加载

1、sklearn模型的保存和加载API

• from sklearn.externals import joblib
  • 保存：joblib.dump(rf, ‘test.pkl’)
  • 加载：estimator = joblib.load(‘test.pkl’)

2、决策树的模型保存加载案例

保存：

import joblib
from sklearn.model_selection import train_test_split
from sklearn.datasets import fetch_20newsgroups, load_iris
from sklearn.tree import DecisionTreeClassifier, export_graphviz

"""
 用决策树对鸢尾花进行分类
 :return:
 """
# 1）获取数据集
iris = load_iris()

# 2）划分数据集
x_train, x_test, y_train, y_test = train_test_split(iris.data, iris.target, random_state=22)

# 3）决策树预估器
estimator = DecisionTreeClassifier(criterion="entropy", )
estimator.fit(x_train, y_train)

joblib.dump(estimator, 'iris.pkl')
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

会生成如下的文件：

加载：

model = joblib.load("iris.pkl")
print(model.predict(x_test))
1
2