The preservation and recovery of machine learning-training model (Sklearn)

When doing model training, especially for cross validation on a training set, you typically want to save the model and then put it on a separate test set, which describes the save and reuse of the training model in Python.

Scikit-learn already has the model to persist the operation, the import joblib can

From sklearn.externals import Joblib

Model Save

>>> 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)
>>> joblib.dump (CLF, " TRAIN_MODEL.M ")

The model can be saved locally by the Joblib dump, and the CLF is a trained classifier model to be recalled from the local

>>> CLF = joblib.load ("TRAIN_MODEL.M")

Load the saved model through the Joblib load method.

And then you can test it on the test set.

Clf.predit (test_x) #此处test_X为特征集