5线性回归算法

print(‘201706120172,罗镕涛,软件1702‘)
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
df = pd.read_csv(‘./data/罗镕涛201706120172-1.csv‘,encoding=‘GB2312‘)
df.iloc[:,5]
X=df.iloc[:,5]#生成-2*3.14到2*3.14的200个数
Y=df.iloc[:,4]-0.1#将X正弦化，然后加入噪音
np.array(X)
np.array(Y)
X=X.values.reshape(-1,1)#扁平化，就是features只能为1
Y=Y.values.reshape(-1,1)
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import PolynomialFeatures
from sklearn.pipeline import Pipeline

def polynomial_model(degree=1):
    polynomial_features=PolynomialFeatures(degree=degree,include_bias=False)#生成degree阶多项式
    linear_regression=LinearRegression(normalize=True)#线性回归实例化，并且正规化
    pipeline=Pipeline([("polynomial_features",polynomial_features),("linear_regression",linear_regression)])#流水线
    return pipeline
from sklearn.metrics import mean_squared_error

degrees=[2,3,5,10]#多项式的阶数
results=[]#结果数组
for d in degrees:#运行四次
    model=polynomial_model(degree=d)#生产degree多项式模型
    model.fit(X,Y)#将X,Y扔到模型里面去训练
    train_score=model.score(X,Y)#得到评分
    mse=mean_squared_error(Y,model.predict(X))#计算均方根误差
    results.append({"model":model,"degree":d,"score":train_score,"mse":mse})#追加对应数据到results里面
for r in results:
    print("degree:{};train score:{};mean squared error:{}".format(r["degree"],r["score"],r["mse"]))
from matplotlib.figure import SubplotParams
plt.figure(figsize=(12,6),dpi=200,subplotpars=SubplotParams(hspace=0.3))
                  #subplotpars只是控制各下属图形高度上的间距为0.3
for i,r in enumerate(results):#循环四次
    fig=plt.subplot(2,2,i+1)#四张图，绘画顺序对应i+1
    plt.xlim(-8,8)#每张图的x轴的限制为-8到8，对应上面的-2*3.14到2*3.14
    plt.title("LinearRegression degree={}".format(r["degree"]))#标题
    plt.scatter(X,Y,s=5,c=‘b‘,alpha=0.5)#蓝色的散点
    plt.plot(X,r["model"].predict(X),‘r-‘)#预测的曲线