glmmTMB拟合Gamma/ziGamma模型时负对数似然NaN错误求助

在使用glmmTMB包拟合Gamma族模型时，遇到如下错误：

Error in fitTMB(TMBStruc) : negative log-likelihood is NaN at starting parameter values

复现代码 #

1. Gamma族模型报错

testdata<-data.frame(DV=1+sample(0:100),
                     pred=sample(-50:50))
hmod<-glmmTMB(formula=DV~pred,
              data=testdata,family=Gamma)

错误信息：

Error in fitTMB(TMBStruc) : 
  negative log-likelihood is NaN at starting parameter values

2. ziGamma模型同样报错

testdata<-data.frame(DV=sample(0:100),
                     pred=sample(-50:50))
hmod<-glmmTMB(formula=DV~pred,
              ziformula=~pred,
              data=testdata,family=ziGamma)

错误信息：

Error in fitTMB(TMBStruc) : 
  negative log-likelihood is NaN at starting parameter values

3. 仅截距项的ziGamma模型（收敛但有NA警告）

hmod<-glmmTMB(formula=DV~1,
              ziformula=~pred,
              data=testdata,family=ziGamma)

警告信息：

Warning messages:
1: In (function (start, objective, gradient = NULL, hessian = NULL,  :
  NA/NaN function evaluation
2: In (function (start, objective, gradient = NULL, hessian = NULL,  :
  NA/NaN function evaluation
3: In (function (start, objective, gradient = NULL, hessian = NULL,  :
  NA/NaN function evaluation
4: In (function (start, objective, gradient = NULL, hessian = NULL,  :
  NA/NaN function evaluation

已尝试的解决方法 #

• 从源码重装了Matrix、TMB和glmmTMB包，问题未解决
• 参考旧帖建议将Matrix降级至1.6-2，但此举会破坏lme4等其他依赖包，因此寻求其他可行解决方案

会话信息 #

sessionInfo()
# R version 4.4.1 (2024-06-14 ucrt)
# Platform: x86_64-w64-mingw32/x64
# Running under: Windows 10 x64 (build 19045)
# 
# Matrix products: default
# 
# 
# locale:
# [1] LC_COLLATE=English_United Kingdom.utf8  LC_CTYPE=English_United Kingdom.utf8   
# [3] LC_MONETARY=English_United Kingdom.utf8 LC_NUMERIC=C                           
# [5] LC_TIME=English_United Kingdom.utf8    
# 
# time zone: Europe/Vienna
# tzcode source: internal
# 
# attached base packages:
# [1] stats     graphics  grDevices utils     datasets  methods   base      
# 
# other attached packages:
#  [1] glmmTMB_1.1.10 lmerTest_3.1-3 lme4_1.1-35.5  Matrix_1.7-1   mgcv_1.9-1     nlme_3.1-164  
#  [7] ggplot2_3.5.1  stringr_1.5.1  dplyr_1.1.4    magrittr_2.0.3
# 
# loaded via a namespace (and not attached):
#  [1] sandwich_3.1-1      utf8_1.2.4          generics_0.1.3      stringi_1.8.4       lattice_0.22-6     
#  [6] grid_4.4.1          estimability_1.5.1  mvtnorm_1.3-2       this.path_2.5.0     survival_3.6-4     
# [11] multcomp_1.4-26     fansi_1.0.6         scales_1.3.0        TH.data_1.1-2       codetools_0.2-20   
# [16] numDeriv_2016.8-1.1 reformulas_0.4.0    Rdpack_2.6.1        cli_3.6.3           rlang_1.1.4        
# [21] rbibutils_2.3       munsell_0.5.1       splines_4.4.1       withr_3.0.2         tools_4.4.1        
# [26] nloptr_2.1.1        coda_0.19-4.1       minqa_1.2.8         colorspace_2.1-1    boot_1.3-30        
# [31] vctrs_0.6.5         R6_2.5.1            zoo_1.8-12          lifecycle_1.0.4     emmeans_1.10.5     
# [36] MASS_7.3-60.2       pkgconfig_2.0.3     pillar_1.9.0        gtable_0.3.6        glue_1.8.0         
# [41] Rcpp_1.0.13-1       tibble_3.2.1        tidyselect_1.2.1    rstudioapi_0.17.1   xtable_1.8-4       
# [46] TMB_1.9.15          compiler_4.4.1 

内容的提问来源于stack exchange，提问作者Arpaxad