python statsmodels 多元线性归回流程结果导入excel

mohana48833985

已于 2024-06-18 11:33:51 修改

阅读量139

点赞数 3

文章标签： python 人工智能机器学习

于 2024-06-06 18:18:01 首次发布

本文链接：https://blog.csdn.net/Caiqiudan/article/details/139507174

版权

import statsmodels.api as sm

train_x['const'] = 1 # 添加截距项
lr = sm.OLS(train_y,train_x)
result = lr.fit()
result.summary2()

# 预测
test_x['const'] = 1
result.predict(test_x)

# 预测结果输出为dataframe，方便保存到excel
a1 = pd.DataFrame(result.summary().tables[0])
a2 = pd.DataFrame(result.summary().tables[1])
df_result = a1.append(a2)
df_result

在这里插入图片描述

模型结果导入excel

# 模型拟合结果
a1 = pd.DataFrame([i.split(',') for i in result.summary().tables[0].as_csv().replace(' ', '').split('\n')])
a2 = pd.DataFrame([i.split(',') for i in result.summary().tables[1].as_csv().replace(' ', '').split('\n')])
df_result = a1.append(a2)
# 增加测试集的误差和训练集的误差
train_res = train_data.误差.describe(percentiles=[0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,0.95,0.97]).rename('训练集误差').reset_index()
test_res = test_data.误差.describe(percentiles=[0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,0.95,0.97]).rename('测试集误差').reset_index()
df_res = pd.merge(train_res, test_res, on=['index'])
df_result = df_res.append(df_result)