序列化和反序列化机器学习的模型
需要将生成的机器学习模型序列化,并将其发布到生产环境。
当有新的数据出现时,需要反序列化已保存的模型,然后用其预测新的数据。
1. 通过pickle序列化和反序列化机器学习模型
pickle是标准的python序列化方法,可以通过它来序列化机器学习算法生成的模型,并将其保存到文件中。当需要对新数据进行预测时,将已保存的模型反序列化,并用其预测新的数据。
# -*- coding: utf-8 -*-
#import matplotlib.pyplot as plt
from pandas import read_csv
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from pickle import dump
from pickle import load
filename='pima indians.txt'
names=['preg','plas','pres','skin','test','mass','pedi','age','class']
data=read_csv(filename,names=names)
array= data.values
X= array[:,0:8]
Y= array[:,8]
test_size = 0.33
seed=4
X_train,X_test, Y_train, Y_test = train_test_split(X,Y, test_size= test_size, random_state =seed)
#训练模型
model = LogisticRegression()
model.fit(X_train,Y_train)
#保存模型
model_file='finalized_model.sav'
with open(model_file, 'wb') as model_f:
#模型序列化
dump(model, model_f)
#加载模型
with open(model_file, 'rb') as model_f:
#模型反序列化
loaded_model= load(model_f)
result=loaded_model.score(X_test, Y_test)
print('算法评估结果:%.3f'%(result*100))
运行结果为:
算法评估结果:80.315
2.通过joblib序列化和反序列化机器学习的模型
joblib是SciPy生态环境的一部分,提供了通用的工具来序列化python的对象和反序列化python的对象。
通过joblib序列化对象时会采用Numpy的格式保存数据,这对某些保存数据到模型中的算法非常有效,如K近邻算法。
# -*- coding: utf-8 -*-
#import matplotlib.pyplot as plt
from pandas import read_csv
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.externals.joblib import dump
from sklearn.externals.joblib import load
filename='pima indians.txt'
names=['preg','plas','pres','skin','test','mass','pedi','age','class']
data=read_csv(filename,names=names)
array= data.values
X= array[:,0:8]
Y= array[:,8]
test_size = 0.33
seed=4
X_train,X_test, Y_train, Y_test = train_test_split(X,Y, test_size= test_size, random_state =seed)
#训练模型
model = LogisticRegression()
model.fit(X_train,Y_train)
#保存模型
model_file='finalized_model.sav'
with open(model_file, 'wb') as model_f:
#模型序列化
dump(model, model_f)
#加载模型
with open(model_file, 'rb') as model_f:
#模型反序列化
loaded_model= load(model_f)
result=loaded_model.score(X_test, Y_test)
print('算法评估结果:%.3f'%(result*100))
运算结果为:
算法评估结果:80.315