[R-sig-ME] specifying initial starting values in glmmTMB

Diego Pavon diego.pavonjordan at gmail.com
Tue Feb 20 13:05:20 CET 2018


Hello,

I am having convergence problems with the following model:

M2 <- glmmTMB(Abundance ~ Temp_std + WithinNatura + Winter_std
              + (1|site) + (1|species),
              family = nbinom2,
              control = glmmTMBControl(optCtrl = list(iter.max = 20000,
eval.max = 20000), profile = TRUE, collect = FALSE),
              ziformula = ~ Temp_std + WithinNatura + Winter_std,
              data = AllSpecies)


I have followed the recommendation in the troubleshooting vignette
(standardize covariates, simplify the model - can't get simpler than the
above-, increase iteractions...) and also in the pdf (
https://cran.r-project.org/web/packages/glmmTMB/glmmTMB.pdf) but it does
not seem to help.

I was wondering if there is a way to specify initial values, perhaps
similar to in lme4?

Thank you very much for your help.

> sessionInfo()
R version 3.3.2 (2016-10-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows >= 8 x64 (build 9200)

locale:
[1] LC_COLLATE=Finnish_Finland.1252  LC_CTYPE=Finnish_Finland.1252
LC_MONETARY=Finnish_Finland.1252 LC_NUMERIC=C
[5] LC_TIME=Finnish_Finland.1252

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

other attached packages:
[1] glmmTMB_0.2.0

loaded via a namespace (and not attached):
 [1] Rcpp_0.12.11    lattice_0.20-35 reshape_0.8.6   unmarked_0.12-2
MASS_7.3-47     grid_3.3.2      plyr_1.8.4      nlme_3.1-131
 [9] minqa_1.2.4     nloptr_1.0.4    raster_2.5-8    sp_1.2-4
Matrix_1.2-10   splines_3.3.2   lme4_1.1-13     tools_3.3.2
[17] TMB_1.7.10      parallel_3.3.2


Best

Diego


-- 
*Diego Pavón Jordán*

*Finnish Museum of Natural History*
*PO BOX 17 *

*Helsinki. Finland*

*0445061210https://www.researchgate.net/profile/Diego_Pavon-jordan
<https://www.researchgate.net/profile/Diego_Pavon-jordan>*

	[[alternative HTML version deleted]]



More information about the R-sig-mixed-models mailing list