使用岭回归对数据进行拟合

• 改变alpha的值对数据进行拟合。

# 岭回归是保留全部的特征变量， 减少特征变量的值来防止过拟合现象
from sklearn.linear_model import Ridge
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split

X, y = load_boston().data, load_boston().target
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=8)
rd = Ridge().fit(X_train, y_train)
print(rd.score(X_train, y_train))
print(rd.score(X_test, y_test))

# 通过增加alpha的值可以降低特征变量的系数
# 增加alpha的值可以降低过拟合程度
# 较高的alpha的值模型的限制级更加严格 coef_的值会更小

from sklearn.datasets import load_diabetes

X, y = load_diabetes().data, load_diabetes().target
X_train, X_test, y_train, y_test = train_test_split(X, y)
rd10 = Ridge(alpha=10).fit(X_train, y_train)
print(rd10.score(X_train, y_train))
print(rd10.score(X_test, y_test))

rd100 = Ridge(alpha=100).fit(X_train, y_train)
print(rd100.score(X_train, y_train))
print(rd100.score(X_test, y_test))

rd01 = Ridge(alpha=0.1).fit(X_train, y_train)
print(rd01.score(X_train, y_train))
print(rd01.score(X_test, y_test))

import matplotlib.pyplot as plt

plt.plot(rd.coef_, '^', label='rd')
plt.plot(rd10.coef_, '*', label='rd10')
plt.plot(rd100.coef_, 'o', label='rd100')
plt.plot(rd01.coef_, 'v', label='rd01')
plt.hlines(0, 0, len(rd.coef_))
plt.legend()
plt.show()

*