python进行数据分析----线性回归

方法：
import statsmodels.api as sm
import pandas as pd
from patsy.highlevel  import dmatrices  ----2.7里面是 from patsy import dmatrices
hg ='D:/hg.csv'
df=pd.read_csv(hg)
vars=['rkzzl','zrs','rjgdp']
df=df[vars]
y,X=dmatrices(' rkzzl ~ zrs + rjgdp ',data=df,return_type='dataframe')
mod=sm.OLS(y,X)
res=mod.fit()
print res.summary()

import statsmodels.api as sm
import pandas as pd
import numpy as np
from patsy.highlevel import dmatrices
from common.util.my_sqlalchemy import sqlalchemy_engine
import math
sql = "select Q1R3, Q1R5, Q1R6, Q1R7 from db2017091115412316222027656281_1;"
df = pd.read_sql(sql, sqlalchemy_engine)
df_dropna = df.dropna()
y,X=dmatrices(' Q1R3 ~ Q1R5 + Q1R6 + Q1R7',data=df_dropna,return_type='dataframe')
mod=sm.OLS(y,X)
res=mod.fit()
result = res.summary()
print(result)
model = {
'n': int(res.nobs),
'df': res.df_model,
'r': math.sqrt(res.rsquared),
'r_squared':res.rsquared,
'r_squared_adj': res.rsquared_adj,
'f_statistic': res.fvalue, # F检验
'prob_f_statistic': res.f_pvalue,
}
coefficient = {
'coefficient':list(res.params),
'std': list(np.diag(np.sqrt(res.cov_params()))),
't': list(res.tvalues),
'sig': [i for i in map(lambda x:float(x),("".join("{:.4f},"*len(res.pvalues)).format(*list(res.pvalues))).rstrip(",").split(","))]
}
returnValue = {'model': model, 'coefficient': coefficient}
print(returnValue)

{
'model': {
'df': 3.0,
'n': 665,
'prob_f_statistic': 1.185607423551511e-17,
'r_squared_adj': 0.11247707470462853,
'f_statistic': 29.049896130483212,
'r_squared': 0.11648696743939679,
'r': 0.3413018714267427},
'coefficient': {
'std': [0.30170364007280126, 0.049972399035516278, 0.051623405028706125, 0.047659986606566104],
'sig': [0.0, 0.0, 0.0, 0.0312],
't': [5.4578212730306044, 5.3469744215460269, 4.3810228293129168, 2.1587543885465008],
'coefficient': [1.6466445449401035, 0.26720113942619689, 0.22616331595762876, 0.10288620524499202]}
}