Error in negative.binomial() : 'theta' must be specified

[caimiao0714]

Hi there,

I was trying to implement the negative binomial model example. However, I got the following error message:

library(GLMMadaptive)
> set.seed(102)
> dd <- expand.grid(f1 = factor(1:3), f2 = LETTERS[1:2], g = 1:30, rep = 1:15,
+                   KEEP.OUT.ATTRS = FALSE)
> mu <- 5 * (-4 + with(dd, as.integer(f1) + 4 * as.numeric(f2)))
> dd$y <- rnbinom(nrow(dd), mu = mu, size = 0.5)
> 
> gm1 <-  mixed_model(fixed = y ~ f1 * f2, random = ~ 1 | g, data = dd,
+                     family = negative.binomial())
Error in negative.binomial() : 'theta' must be specified

here is my session information:

> sessionInfo()
R version 3.6.0 (2019-04-26)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Red Hat Enterprise Linux

Matrix products: default
BLAS/LAPACK: /opt/intel/18.0.3/compilers_and_libraries_2018.3.222/linux/mkl/lib/intel64_lin/libmkl_gf_lp64.so

locale:
 [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C               LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
 [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=en_US.UTF-8    LC_PAPER=en_US.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C             LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] parallel  stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] optimParallel_0.8-1 dplyr_0.8.3         data.table_1.12.8   nloptr_1.2.1        optimx_2018-7.10   
[6] lme4_1.1-21         Matrix_1.2-17       GLMMadaptive_0.6-9 

loaded via a namespace (and not attached):
 [1] tidyselect_0.2.5    remotes_2.1.0       purrr_0.3.2         splines_3.6.0       lattice_0.20-38    
 [6] testthat_2.2.1      usethis_1.5.1       rlang_0.4.0         pkgbuild_1.0.5      pillar_1.4.2       
[11] glue_1.3.1          withr_2.1.2         sessioninfo_1.1.1   matrixStats_0.55.0  devtools_2.1.0     
[16] memoise_1.1.0       fst_0.9.0           callr_3.3.1         ps_1.3.0            curl_4.0           
[21] Rcpp_1.0.3          backports_1.1.4     desc_1.2.0          pkgload_1.0.2       fs_1.3.1           
[26] packrat_0.5.0       digest_0.6.20       processx_3.4.1      numDeriv_2016.8-1.1 grid_3.6.0         
[31] rprojroot_1.3-2     cli_1.1.0           tools_3.6.0         magrittr_1.5        tibble_2.1.3       
[36] pacman_0.5.1        crayon_1.3.4        pkgconfig_2.0.2     MASS_7.3-51.4       prettyunits_1.0.2  
[41] assertthat_0.2.1    minqa_1.2.4         rstudioapi_0.10     R6_2.4.0            boot_1.3-22        
[46] nlme_3.1-139        compiler_3.6.0

Thank you, Miao

[drizopoulos]

This logically happens because the negative.binomial() family function from lme4 has overwritten the function with the same name from GLMMadaptive. Try fitting again the model by specifying in mixed_model() the family argument as family = GLMMadaptive:::negative.binomial().

[caimiao0714]

This logically happens because the negative.binomial() family function from lme4 has overwritten the function with the same name from GLMMadaptive. Try fitting again the model by specifying in mixed_model() the family argument as family = GLMMadaptive:::negative.binomial().

Thank you! Is it possible to specify an offset term in mixed-effect negative binomial models?

[drizopoulos]

Yes, you can include an offset directly into the fixed effects formula.