比较python中的混合模型

比较python中的混合模型,python,r,statsmodels,nlme,Python,R,Statsmodels,Nlme,我试图使用R的nlme包来比较两个模型 import rpy2.robjects.robject as r >> politeness = pd.read_csv('http://www.bodowinter.com/tutorial/politeness_data.csv') >> mdl1=nlme.gls(Formula('frequency ~ 1'), data=politeness, method="ML", na_action="na.omit") >

我试图使用
R
nlme
包来比较两个模型

import rpy2.robjects.robject as r

>> politeness = pd.read_csv('http://www.bodowinter.com/tutorial/politeness_data.csv')
>> mdl1=nlme.gls(Formula('frequency ~ 1'), data=politeness, method="ML", na_action="na.omit")
>> mdl2=nlme.lme(Formula('frequency ~ 1'), data=politeness, method="ML", random=Formula("~1|subject"), na_action="na.omit")
>> print(r.anova(mdl1,mdl2))
这打印了很多输出,但不是我真正感兴趣的输出。在
R
中,我只得到:

  Model df      AIC      BIC    logLik   Test  L.Ratio p-value
mdl1     1  2 932.8611 937.6988 -464.4306                        
mdl2     2  3 833.2497 840.5063 -413.6249 1 vs 2 101.6114  <.0001
模型df AIC BIC logLik试验L.比率p值
mdl1 1 2 932.8611 937.6988-464.4306
mdl2 2 3 833.2497840.5063-413.6249 1对2 101.6114