机器学习保存和调用模型

以泰坦尼克号数据集（Logistic回归模型为例）

from sklearn import model_selection
from sklearn import linear_model
**import pickle** **导入pickle模块**
predictor = ttnk.columns[1:]
X = ttnk[predictor]
y = ttnk.Survived
X_train,X_test,y_train,y_test = model_selection.train_test_split(X,y,test_size = 0.25,random_state = 1234)

model = linear_model.LogisticRegression()
model.fit(X_train,y_train)
predict_data = model.predict(X_test)

***s = pickle.dumps(model)
    model2 = pickle.loads(s)*** 很关键的两步pickle.dumps（保存模型）                                                pickle.loads(调用模型）
print(df)