1.Sklearn框架KNN算法回归篇

# 导包
from sklearn.neighbors import KNeighborsRegressor
import numpy as np
import matplotlib.pyplot as plt


# 生成样本数据
x=np.linspace(-np.pi,np.pi,40)
# display(x)
y=np.sin(x)
display(y)
plt.scatter(x,y)

# 增加噪音数据  [0,1)  [-0.5,0.5)
noise=np.random.random(size=20)-0.5
# noise


# 将y中的40个值每隔一个数值加上噪音数据，也就是让奇数位随机的在-0.5到0.5之间浮动
y[::2]+=noise


plt.scatter(x,y)

# 生成测试数据
x_train=x.reshape(-1,1)
y_train=y
display(x_train,y)


# 构建KNN对象   训练数据
knn=KNeighborsRegressor(n_neighbors=7)
knn.fit(x_train,y_train)


# 测试数据
x_test=np.linspace(-np.pi,np.pi,100).reshape(-1,1)
x_test
# 开始预测
y_=knn.predict(x_test)


plt.plot(x_test,y_,color="red")
plt.scatter(x_train,y_train,color="green")

# 欠拟合  过拟合