glmer模型的残差方差丢失

时间:2016-12-28 15:11:20

标签: r lme4

我正在尝试为glmer()lme4包中的R随机拦截模型计算intraclass correlation

以下是模型的外观:

model1 <- lme4::glmer(funded ~ 
                          tweeted +
                          `Project Resource Type` + 
                          `Project Subject Category` +
                          `School Poverty Level` +
                      (1 | `School State`),
                      family = "binomial",
                      data = df)

这是输出(省略固定效果):

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: binomial  ( logit )
Formula: funded ~ tweeted + `Project Resource Type` + `Project Subject Category` +  
    `School Poverty Level` + (1 | `School State`)
   Data: df

      AIC       BIC    logLik  deviance  df.resid 
 599914.4  600091.2 -299941.2  599882.4    465409 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-2.2563 -1.2068  0.6538  0.7744  1.3985 

Random effects:
 Groups       Name        Variance Std.Dev.
 School State (Intercept) 0.04272  0.2067  
Number of obs: 465425, groups:  School State, 51

Random effects:下面,我预计会有个人(在这种情况下,来自Donors ChooseResidual VarianceStd.Dev.估算的单个项目,以及小组(School StateVarianceStd.Dev.估算,但没有。

可以从输出中提取这些估算值吗?有没有理由不包括在内?我知道这可能是一个与Cross Validated相关的以统计为中心的问题,但我认为因为它与输出有关,这里是合适的。

0 个答案:

没有答案