在 R 中运行 glmer 时出错：特征值比大 - 重缩放变量

问：

我收到一个奇怪的警告，在运行具有二元结果的广义线性混合模型时，我不明白。

型号代码

e = glmer(Sexb ~ `Age class` + `Hunting method def` + group_d + `Age class`:group_d + (1 | `Hunting season`),
          data = modeldata, 
          family = binomial
          )

Warning: Model is nearly unidentifiable: large eigenvalue ratio
 - Rescale variables?

模型摘要

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) ['glmerMod']
 Family: binomial  ( logit )
Formula: Sexb ~ `Age class` + `Hunting method def` + group_d + `Age class`:group_d +      (1 | `Hunting season`)
   Data: modeldata

     AIC      BIC   logLik deviance df.resid 
  2618.3   2696.5  -1295.2   2590.3     1943 

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-4.9826 -0.8979 -0.6678  1.0882  1.6015 

Random effects:
 Groups         Name        Variance Std.Dev.
 Hunting season (Intercept) 0.01488  0.122   
Number of obs: 1957, groups:  Hunting season, 6

Fixed effects:
                                  Estimate Std. Error z value Pr(>|z|)  
(Intercept)                        0.33915    0.26018   1.304   0.1924  
`Age class`1                      -0.99888    0.90146  -1.108   0.2678  
`Age class`2                       0.24938    0.67717   0.368   0.7127  
`Age class`3+                      1.08310    0.52225   2.074   0.0381 *
`Hunting method def`HD            -0.08716    0.13083  -0.666   0.5053  
group_dFamily Pack                -0.37231    0.30945  -1.203   0.2289  
group_dSolo                       -0.70229    0.28811  -2.438   0.0148 *
`Age class`1:group_dFamily Pack   21.30707  126.26487   0.169   0.8660  
`Age class`2:group_dFamily Pack  -17.77823   66.55510  -0.267   0.7894  
`Age class`3+:group_dFamily Pack   2.22920    1.15709   1.927   0.0540 .
`Age class`1:group_dSolo           1.21386    0.91569   1.326   0.1850  
`Age class`2:group_dSolo          -0.62667    0.69968  -0.896   0.3704  
`Age class`3+:group_dSolo         -0.89530    0.54769  -1.635   0.1021  
---
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Correlation matrix not shown by default, as p = 13 > 12.
Use print(x, correlation=TRUE)  or
    vcov(x)        if you need it

optimizer (Nelder_Mead) convergence code: 0 (OK)
Model is nearly unidentifiable: large eigenvalue ratio
 - Rescale variables?

关于数据的几句话

我的数据包含 5 个变量：

• 狩猎方法（HD 或 DD）
• 性别（男性或女性）=>转换为性别b，其中 1 = 女性
• 年龄级别 （0,1,2,3+）
• 狩猎季节（6季）
• Group_d （Sexb）

Summary of the data

 Hunting method def     Sex       Age class Hunting season        group_d     Sexb    
 DD:1657            female: 867   0 :416    11_12:338      Duo        : 111   0:1090  
 HD: 300            male  :1090   1 :599    12_13:341      Family Pack: 162   1: 867  
                                  2 :388    13_14:433      Solo       :1684           
                                  3+:554    14_15:408                                 
                                            17_18:287                                 
                                            20_21:150                                 

由于我对 glmer 没有经验，我真的不知道如何处理警告。此源 （https://rstudio-pubs-static.s3.amazonaws.com/33653_57fc7b8e5d484c909b615d8633c01d51.html） 给出了重新缩放连续变量的解决方案。但是，由于所有可能的变量都是因素，我不知道我应该如何应用它。

如何处理此警告？