python 回归分析

来源：互联网发布：sql 统计每小时记录编辑：程序博客网时间：2024/06/05 19:33

pwd

‘d:\\python\\exerise-df\\df-data-analysis’

from scipy import statsimport pandas as pdimport numpy as npfrom statsmodels.formula.api import olsimport statsmodels.api as smfrom statsmodels.stats.anova import anova_lmfrom statsmodels.stats.multicomp import pairwise_tukeyhsdimport matplotlib.pyplot as plt

单变量分析分析

dat = pd.read_csv("simple-resgreesion.csv")

dat.head()

N weight 0 58 115 1 59 117 2 60 120 3 61 123 4 62 126

model = ols('weight ~ N',dat).fit()

print(model.summary())

                            OLS Regression Results                            ==============================================================================Dep. Variable:                 weight   R-squared:                       0.991Model:                            OLS   Adj. R-squared:                  0.990Method:                 Least Squares   F-statistic:                     1433.Date:                Wed, 27 Sep 2017   Prob (F-statistic):           1.09e-14Time:                        14:49:40   Log-Likelihood:                -26.541No. Observations:                  15   AIC:                             57.08Df Residuals:                      13   BIC:                             58.50Df Model:                           1                                         Covariance Type:            nonrobust                                         ==============================================================================                 coef    std err          t      P>|t|      [95.0% Conf. Int.]------------------------------------------------------------------------------Intercept    -87.5167      5.937    -14.741      0.000      -100.343   -74.691N              3.4500      0.091     37.855      0.000         3.253     3.647==============================================================================Omnibus:                        2.396   Durbin-Watson:                   0.315Prob(Omnibus):                  0.302   Jarque-Bera (JB):                1.660Skew:                           0.789   Prob(JB):                        0.436Kurtosis:                       2.596   Cond. No.                         982.==============================================================================Warnings:[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.

多项式回归分析

dat2 = pd.read_csv("duoxiangshi.csv")

dat2.head()

N weight 0 58 115 1 59 117 2 60 120 3 61 123 4 62 126

mod = ols('weight ~ N + I(N**2)',dat2).fit()

print(mod.summary())

                            OLS Regression Results                            ==============================================================================Dep. Variable:                 weight   R-squared:                       0.999Model:                            OLS   Adj. R-squared:                  0.999Method:                 Least Squares   F-statistic:                 1.139e+04Date:                Wed, 27 Sep 2017   Prob (F-statistic):           2.13e-20Time:                        14:59:57   Log-Likelihood:                -5.2563No. Observations:                  15   AIC:                             16.51Df Residuals:                      12   BIC:                             18.64Df Model:                           2                                         Covariance Type:            nonrobust                                         ==============================================================================                 coef    std err          t      P>|t|      [95.0% Conf. Int.]------------------------------------------------------------------------------Intercept    261.8782     25.197     10.393      0.000       206.979   316.777N             -7.3483      0.778     -9.449      0.000        -9.043    -5.654I(N ** 2)      0.0831      0.006     13.891      0.000         0.070     0.096==============================================================================Omnibus:                        2.449   Durbin-Watson:                   1.144Prob(Omnibus):                  0.294   Jarque-Bera (JB):                1.033Skew:                           0.049   Prob(JB):                        0.597Kurtosis:                       1.718   Cond. No.                     1.09e+06==============================================================================

多变量回归分析

dat = pd.read_csv("mul-regression.csv")

dat.head()

x1 x2 x3 x4 y 0 30.8 33.0 50.0 90 520.8 1 23.6 33.6 28.0 64 195.0 2 31.5 34.0 36.6 82 424.0 3 19.8 32.0 36.0 70 213.5 4 27.7 26.0 47.2 74 403.3

mod = ols('y ~ x1 + x2 + x3 + x4',dat).fit()

print(mod.summary())

                            OLS Regression Results                            ==============================================================================Dep. Variable:                      y   R-squared:                       0.894Model:                            OLS   Adj. R-squared:                  0.866Method:                 Least Squares   F-statistic:                     31.78Date:                Wed, 27 Sep 2017   Prob (F-statistic):           3.66e-07Time:                        14:52:33   Log-Likelihood:                -97.454No. Observations:                  20   AIC:                             204.9Df Residuals:                      15   BIC:                             209.9Df Model:                           4                                         Covariance Type:            nonrobust                                         ==============================================================================                 coef    std err          t      P>|t|      [95.0% Conf. Int.]------------------------------------------------------------------------------Intercept   -625.3583    114.378     -5.467      0.000      -869.150  -381.566x1            15.1962      2.127      7.146      0.000        10.663    19.729x2             7.3785      1.889      3.907      0.001         3.353    11.404x3             9.5034      1.342      7.082      0.000         6.643    12.364x4            -0.8468      1.493     -0.567      0.579        -4.029     2.335==============================================================================Omnibus:                        0.492   Durbin-Watson:                   1.620Prob(Omnibus):                  0.782   Jarque-Bera (JB):                0.578Skew:                          -0.294   Prob(JB):                        0.749Kurtosis:                       2.409   Cond. No.                     1.38e+03==============================================================================

阅读全文

0 0