我正在尝试使用python 3.4在statsmodels 0.6模块中使用负二项分析数据集。
import pandas as pd
import statsmodels.formula.api as smf
import statsmodels.api as sm
model2=smf.glm("an_darlingi ~ trat", data=dados,
family=sm.families.NegativeBinomial()).fit()
输出：
Generalized Linear Model Regression Results
==============================================================================
Dep. Variable: an_darlingi No. Observations: 363
Model: GLM Df Residuals: 352
Model Family: NegativeBinomial Df Model: 10
Link Function: log Scale: 1.52662855632
Method: IRLS Log-Likelihood: -1404.7
Date: Mon, 13 Oct 2014 Deviance: 578.37
Time: 19:34:45 Pearson chi2: 537.
No. Iterations: 7
=================================================================================
coef std err t P>|t| [95.0% Conf. Int.]
---------------------------------------------------------------------------------
Intercept 2.9988 0.220 13.607 0.000 2.567 3.431
trat[T.CO2] -0.3070 0.313 -0.981 0.327 -0.920 0.306
trat[T.ML] -0.1709 0.312 -0.547 0.584 -0.783 0.441
trat[T.SFPF] -0.3215 0.313 -1.027 0.304 -0.935 0.292
trat[T.SFPM] -0.4085 0.314 -1.303 0.193 -1.023 0.206
trat[T.SFTF] -0.0448 0.312 -0.144 0.886 -0.656 0.566
trat[T.SFTM] -0.1835 0.312 -0.587 0.557 -0.796 0.429
trat[T.SIPF] -0.3905 0.313 -1.246 0.213 -1.005 0.224
trat[T.SIPM] -0.1799 0.312 -0.576 0.565 -0.792 0.432
trat[T.SIPTM] 0.0700 0.311 0.225 0.822 -0.540 0.680
trat[T.SITF] 0.1968 0.311 0.633 0.527 -0.413 0.806
=================================================================================
在输出中，我得到了t和p的值，我喜欢得到z的值和P＆gt;（z），就像在R中一样。我怎么能用smf.glm代替sm.NegativeBinomial。因为如果我尝试使用sm.NegativeBinomial，我会收到错误信息，因为它无法将字符串SIPTM转换为浮点数。