https://www.lanqiao.cn/courses/

岭回归_最小二乘法的改良版，有L2范数惩罚函数

from sklearn.linear_model import Ridge #Lasso回归_有L1范数惩罚函数 from sklearn.linear_model import Lasso #贝叶斯岭回归 from sklearn.linear_model import BayesianRidge #随机梯度下降回归 from sklearn.linear_model import SGDRegressor #鲁棒回归 from sklearn.linear_model import HuberRegressor

导入模型

model_Ridge = Ridge() model_Lasso = Lasso() model_BayesianRidge = BayesianRidge() model_SGDRegressor = SGDRegressor() model_HuberRegressor = HuberRegressor() model_LinearRegression = LinearRegression()

训练模型，利用前面的train/test集

用训练集进行训练模型

model_Ridge.fit(train_feature,train_target) model_Lasso.fit(train_feature,train_target) model_BayesianRidge.fit(train_feature,train_target) model_SGDRegressor.fit(train_feature,train_target) model_HuberRegressor.fit(train_feature,train_target) model_LinearRegression.fit(train_feature,train_target)

输出各个模型的常数项和拟合直线w项

print('model_Ridge:intercept:{},coef:{}'.format(model_Ridge.intercept_,model_Ridge.coef_)) print('model_Lasso:intercept:{},coef:{}'.format(model_Lasso.intercept_,model_Lasso.coef_)) print('model_BayesianRidge:intercept:{},coef:{}'.format(model_BayesianRidge.intercept_,model_BayesianRidge.coef_)) print('model_SGDRegressor:intercept:{},coef:{}'.format(model_SGDRegressor.intercept_,model_SGDRegressor.coef_)) print('model_HuberRegressor:intercept:{},coef:{}'.format(model_HuberRegressor.intercept_,model_HuberRegressor.coef_)) print('model_LinearRegression:intercept:{},coef:{}'.format(model_LinearRegression.intercept_,model_LinearRegression.coef_))

用测试集进行测试并将结果绘图

fig,axes = plt.subplots(2,3,figsize=(16,7)) #岭回归 axes[0,0].scatter(train_feature,train_target,color='red') axes[0,0].scatter(test_feature,test_target,color='yellow') axes[0,0].plot(test_feature, model_Ridge.predict(test_feature),color='black',linewidth=2) axes[0,0].set_title('Ridge') axes[0,0].legend(('Train Set','Test Set','Ridge line'),loc=4) #Lasso回归 axes[0,1].scatter(train_feature,train_target,color='red') axes[0,1].scatter(test_feature,test_target,color='yellow') axes[0,1].plot(test_feature, model_Lasso.predict(test_feature),color='green',linewidth=2) axes[0,1].set_title('Lasso') axes[0,1].legend(('Train Set','Test Set','Lasso line'),loc=4) #贝叶斯岭回归 axes[0,2].scatter(train_feature,train_target,color='red') axes[0,2].scatter(test_feature,test_target,color='yellow') axes[0,2].plot(test_feature, model_BayesianRidge.predict(test_feature),color='purple',linewidth=2) axes[0,2].set_title('BayesianRidge') axes[0,2].legend(('Train Set','Test Set','BayesianRidge line'),loc=4)

axes[1,0].scatter(train_feature,train_target,color='red') axes[1,0].scatter(test_feature,test_target,color='yellow') axes[1,0].plot(test_feature, model_SGDRegressor.predict(test_feature),color='white',linewidth=2) axes[1,0].set_title('SGDRegressor') axes[1,0].legend(('Train Set','Test Set','SGDRegressor line'),loc=4)

axes[1,1].scatter(train_feature,train_target,color='red') axes[1,1].scatter(test_feature,test_target,color='yellow') axes[1,1].plot(test_feature, model_HuberRegressor.predict(test_feature),color='gray',linewidth=2) axes[1,1].set_title('HuberRegressor') axes[1,1].legend(('Train Set','Test Set','HuberRegressor line'),loc=4)

axes[1,2].scatter(train_feature,train_target,color='red') axes[1,2].scatter(test_feature,test_target,color='yellow') axes[1,2].plot(test_feature, model_LinearRegression.predict(test_feature),color='cyan',linewidth=2) axes[1,2].set_title('LinearRegression') axes[1,2].legend(('Train Set','Test Set','LinearRegression line'),loc=4)

理解了监督学习和非监督学习的概念和区别，以及回归问题和分类问题的区别和联系。

哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈

哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈哈

hello world

了解了除 K-Means 之外的其他几种聚类方法

机器学习调参数之路就此开始

简单

11111111

暂停一下。

其实有时候就是需要从最基础的开始，才能慢慢理解。

不能计算时间。

是无法启动环境了吗，您可以在讨论区详细说明一下

实验环境怎么突然用不了了，扎心了。