python机器学习之线性回归详解

#OLS线性回归
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
%matplotlib inline
data = pd.DataFrame(pd.read_excel(r'C:/Users/15643/Desktop/附件1.xlsx'))
feature_data = data.drop(['企业信誉评估'],axis=1)
target_data = data['企业信誉评估']
X_train,X_test,y_train, y_test = train_test_split(feature_data, target_data, test_size=0.3)
from statsmodels.formula.api import ols
from statsmodels.sandbox.regression.predstd import wls_prediction_std
df_train = pd.concat([X_train,y_train],axis=1)
lr_model = ols("企业信誉评估~销项季度均值+有效发票比例+是否违约+企业供求关系+行业信誉度+销项季度标准差",data=df_train).fit()
print(lr_model.summary())
# 预测测试集
lr_model.predict(X_test)

from linear_regression import GDLinearRegression
gd_lr = GDLinearRegression(n_iter=3000,eta=0.001,tol=0.00001)
#梯度下降最大迭代次数n_iter
#学习率eta
#损失降低阈值tol

#scikit-learn 多项式拟合(多元多项式回归)
#PolynomialFeatures和linear_model的组合 (线性拟合非线性)
#[x1,x2,x3]==[[1,x1,x1**2],[1,x2,x2**2],[1,x3,x3**2]]
import numpy as np
import matplotlib.pyplot as plt
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression,Perceptron
from sklearn.metrics import mean_squared_error,r2_score
from sklearn.model_selection import train_test_split
target = std_df_female['总分']
data_complete_ = std_df_female.loc[:,['1000/800','50m','立定跳远','引仰']]
x_train, x_test, y_train, y_test = train_test_split(data_complete_,target, test_size=0.3)
# 多项式拟合
poly_reg  =PolynomialFeatures(degree=2)
x_train_poly = poly_reg.fit_transform(x_train)
model = LinearRegression()
model.fit(x_train_poly, y_train)
#print(poly_reg.coef_,poly_reg.intercept_) #系数及常数
# 测试集比较
x_test_poly = poly_reg.fit_transform(x_test)
y_test_pred = model.predict(x_test_poly)
#mean_squared_error(y_true, y_pred) #均方误差回归损失,越小越好。
mse = np.sqrt(mean_squared_error(y_test, y_test_pred))
# r2 范围[0，1]，R2越接近1拟合越好。
r2 = r2_score(y_test, y_test_pred)
print(r2)