其他回归模型

线性回归 

import numpy as np
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_squred_error, r2_score
from sklearn import datasets
from sklearn.linear_model import LinearRegression, Ridge, Lasso, ElasticNet, ElasticNetCV, LassoCV

diabetes = datasets.load_diabetes()
X = diabetes['data']
y = diabetes['target']

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.15)

lr = LinearRegression()
lr.fit(X_train, y_train)
# 回归问题的得分，不是准确率
lr.score(X_test, y_test)
# 0.5243489738860714

‘’‘
关于lr.score()的解释
The coefficient R^2 is defined as (1 - u/v), where u is the residual
sum of squares ((y_true - y_pred) ** 2).sum() and v is the total
sum of squares ((y_true - y_true.mean()) ** 2).sum().
The best possible score is 1.0 and it can be negative (because the
model can be arbitrarily worse). A constant model that always
predicts the expected value of y, disregarding the input features,
would get a R^2 score of 0.0.
’‘’

# 验证一下
y_ = lr.predict(y_test)
u = ((y_test - y_)**2).sum()
v = ((y_test - y_test.mean()) ** 2).sum()
r2 = 1-u/v
# 0.5243489738860714, 结果和上面lr.score()出来的结果一致

# 用r2_score()验证
r2_score(y_test, y_)
# 0.5243489738860714, 结果也是一样的

mean_squared_error(y_test, y_)
# 2363.0603637069144，mse值越小越好

岭回归

ridge = Ridge(alpha=0.001)
ridge.fit(X_train, y_train)
print(ridge.score(X_test, y_test))
# 0.5233570793503599
y_ = ridge.predict(X_test)
mean_squared_error(y_test, y_)
# 2367.9881501167683

# 对比上面的线性回归
lr = LinearRegression()
lr.fit(X_train,y_train)
print(lr.score(X_test,y_test))
# 0.5243489738860714
y_ = lr.predict(X_test)
mean_squared_error(y_test,y_)
# 2363.0603637069144

# 可以看到，对于特征值之间没有共线性关系的简单数据来说，线性回归模型的效果可能比岭回归的效果更好，但是对于复杂数据来说，一定是岭回归的效果要好一些，而且岭回归模型还要受到alpha值选取的影响

# 使用交叉验证来寻找更优的alpha值
from sklearn.linear_model import RidgeCV
ridgeCV = RidgeCV(alphas=np.logspace(-5, 1, 50), scoring='r2', cv=6)
# 说明一下np.logspace的用法，-5和1分别表示从log10(10^-5)开始到log10(10^1)结束（也就是从0.00001到1之间），取之间的50个数，取得比较均匀，
ridgeCV.fit(X_train, y_train)
y_ = ridgeCV.predict(X_test)
r2_score(y_test, y_)
# 0.5151565371265989, 这个结果貌似还不如上面取0.001时的效果，说明取得仍然不够紧凑

其他线性回归模型

 最小二乘法

岭回归

 Lasso

Multi-task Lasso

 Elastic-Net

 Logistic regression