sklearn 模型保存

from sklearn.externals import joblib

模型训练 

os.chdir("workspace/model_save")
from sklearn import svm
X = [[0, 0], [1, 1]]
y = [0, 1]
clf = svm.SVC()
clf.fit(X, y)  
clf.fit(train_X,train_y)

模型保存

save_path_name=model_save_path+"svm_"+"train_model.m"
self.is_exist(model_save_path,save_path_name)
joblib.dump(clf, save_path_name)
clf = joblib.load(save_path_name)

joblib.dump(clf, "train_model.m")

模型从本地调回

clf = joblib.load("train_model.m")

通过joblib的load方法，加载保存的模型。

然后就可以在测试集上测试了

模型预测

clf.predit(test_X) #此处test_X为特征集