zoukankan      html  css  js  c++  java
  • 机器学习模型问题

    #划分train
    from sklearn.model_selection import train_test_split 
    X_train, X_test, y_train, y_test = train_test_split(train, target, random_state=1)
    
    #构造Ridge的模型
    from sklearn import linear_model
    lassoRegression = linear_model.Ridge()
    lassoRegression.fit(X_train, y_train)
    print("权重向量:%s, b的值为:%.2f" % (lassoRegression.coef_, lassoRegression.intercept_))
    print("损失函数的值:%.2f" % np.mean((lassoRegression.predict(X_test) - y_test) ** 2))
    print("预测性能得分: %.2f" % lassoRegression.score(X_test, y_test))
    
    
    #构造Lasso的模型
    ridgeRegression = linear_model.Lasso()
    ridgeRegression.fit(X_train, y_train)
    print("权重向量:%s, b的值为:%.2f" % (ridgeRegression.coef_, ridgeRegression.intercept_))
    print("损失函数的值:%.2f" % np.mean((ridgeRegression.predict(X_test) - y_test) ** 2))
    print("预测性能得分: %.2f" % ridgeRegression.score(X_test, y_test))
    
     
    
    
    
    #Lasso预测
    y_pred = lassoRegression.predict(X_test)
    
    #Ridge预测
    y_pred = ridgeRegression.predict(X_test)
    
    
    
    #mean_squared_error值
    msn = np.sqrt(mean_squared_error(y_test, y_pred))
    mean_squared_error越小越好
  • 相关阅读:
    Java 类和对象12
    Java类和对象11
    java 类和对象10
    Java类和对象9
    Java类和对象8
    Java 类和对象7
    包装与方法
    JAVA链表
    泛型
    多态 接口
  • 原文地址:https://www.cnblogs.com/wzwi/p/10862120.html
Copyright © 2011-2022 走看看