[R-sig-Geo] Error in splm random effects ML model

Wed Feb 15 19:51:49 CET 2023

Dear Roger, 

Yes, I am running R version 4.2.2 and Windows 10 OS x64, packages used updated to the latest version on CRAN. The panel data I am using is complete for all the variables and spatial weights, which made me wonder why the "within" specification runs perfectly fine, while the "random" one breaks down.

Here is the error traceback:

11: .C64("aplsb1", SIGNATURE = c(SS$signature, SS$signature, "double",
        SS$signature, SS$signature, "double", "double", SS$signature,
        SS$signature, "double", SS$signature, SS$signature, SS$signature,
        SS$signature), nrow, ncol, A using entries, A using colindices, A using rowpointers,
        s, B using entries, B using colindices, B using rowpointers, entries = vector_dc("double",
            nzmax), colindices = vector_dc(SS$type, nzmax), rowpointers = vector_dc(SS$type,
            nrow + 1), nzmax + 1, ierr = vector_dc(SS$type, 1), INTENT = c("r",
            "r", "r", "r", "r", "r", "r", "r", "r", "w", "w", "w",
            "r", "w"), NAOK = getOption("spam.NAOK"), PACKAGE = SS$package)
10: spam_add(e1, e2, -1)
9: I - lambda * csrw
8: I - lambda * csrw
7: xprodB(lambda, w)
6: invSigma(philambda, n, t., w2)
5: objective(.par, ...)
4: nlminb(start = myparms0, objective = ll.c, gradient = NULL, hessian = NULL,
       y = y, X = X, n = n, t. = t., w = w, w2 = w2, scale = parscale,
       control = list(x.tol = x.tol, rel.tol = rel.tol, trace = trace),
       lower = lower.bounds, upper = upper.bounds)
3: est.fun(X, y, ind, tind, n, k, t, nT, w = w, w2 = w2, coef0 = coef0,
       hess = hess, trace = trace, x.tol = x.tol, rel.tol = rel.tol,
       ...)
2: spreml(formula = formula, data = data, index = index, w = listw2mat(listw),
       w2 = listw2mat(listw2), lag = lag, errors = errors, cl = cl,
       ...)
1: spml(formula = mx.standard ~ I.score, data = analysis.subset,
       listw = listw.wgts, effect = "individual", model = "random",
       lag = FALSE, spatial.error = "b", index = c("Group", "Year"))

My sessionInfo:

R version 4.2.2 (2022-10-31 ucrt)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 19044)

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.utf8  LC_CTYPE=English_United States.utf8    LC_MONETARY=English_United States.utf8
[4] LC_NUMERIC=C                           LC_TIME=English_United States.utf8    

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

other attached packages:
 [1] splm_1.6-2      plm_2.6-2       spdep_1.2-7     spData_2.2.1    sp_1.6-0        forcats_1.0.0   stringr_1.5.0   dplyr_1.1.0     purrr_1.0.1    
[10] readr_2.1.4     tidyr_1.3.0     tibble_3.1.8    ggplot2_3.4.1   tidyverse_1.3.2 sf_1.0-9       

loaded via a namespace (and not attached):
 [1] nlme_3.1-160        fs_1.6.1            spatialreg_1.2-6    lubridate_1.9.2     httr_1.4.4          tools_4.2.2         backports_1.4.1    
 [8] utf8_1.2.3          R6_2.5.1            KernSmooth_2.23-20  DBI_1.1.3           colorspace_2.1-0    withr_2.5.0         tidyselect_1.2.0   
[15] compiler_4.2.2      cli_3.6.0           rvest_1.0.3         expm_0.999-7        xml2_1.3.3          sandwich_3.0-2      scales_1.2.1       
[22] lmtest_0.9-40       classInt_0.4-8      proxy_0.4-27        digest_0.6.31       pkgconfig_2.0.3     dbplyr_2.3.0        collapse_1.9.2     
[29] ibdreg_0.3.8        rlang_1.0.6         readxl_1.4.2        rstudioapi_0.14     generics_0.1.3      zoo_1.8-11          jsonlite_1.8.4     
[36] googlesheets4_1.0.1 magrittr_2.0.3      Formula_1.2-4       s2_1.1.2            dotCall64_1.0-2     Matrix_1.5-3        Rcpp_1.0.10        
[43] munsell_0.5.0       fansi_1.0.4         lifecycle_1.0.3     stringi_1.7.8       MASS_7.3-58.1       grid_4.2.2          parallel_4.2.2     
[50] bdsmatrix_1.3-6     crayon_1.5.2        deldir_1.0-6        lattice_0.20-45     haven_2.5.1         splines_4.2.2       hms_1.1.2          
[57] pillar_1.8.1        boot_1.3-28         LearnBayes_2.15.1   wk_0.7.1            reprex_2.0.2        glue_1.6.2          modelr_0.1.10      
[64] vctrs_0.5.2         spam_2.9-1          tzdb_0.3.0          Rdpack_2.4          miscTools_0.6-26    cellranger_1.1.0    gtable_0.3.1       
[71] assertthat_0.2.1    rbibutils_2.2.13    broom_1.0.3         e1071_1.7-13        coda_0.19-4         class_7.3-20        googledrive_2.0.0  
[78] gargle_1.3.0        units_0.8-1         maxLik_1.5-2        timechange_0.2.0    ellipsis_0.3.2     

I tried running the same code on Linux x64 platform as well, with much the same end result.
R version 4.2.2 Patched (2022-11-10 r83330)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 18.04.6 LTS

Here is the link to the minimally reproducible example (code and the actual data):
https://www.dropbox.com/sh/s680xpvfdxowuor/AAAH5yF4JtrzWUgdbvQIF6LBa?dl=0

Thank you for your help.

Denys Dukhovnov

On Wednesday, February 15, 2023 at 11:12:31 AM EST, Roger Bivand <roger.bivand using nhh.no> wrote: 

On Wed, 15 Feb 2023, Denys Dukhovnov via R-sig-Geo wrote:

> Dear community,
>
> I am trying to run a spatial panel random effects SEM model using splm 
> package on a balanced spatial panel of N = 2472 and T = 4, and weights 
> are based on k = 6 nearest neighbors. I encounter this error, which I 
> can't find anything about in the documentation or online.
>
> spml(formula = mx.standard ~ I.score, 
>                          data = analysis.subset,
>                          listw = listw.wgts,
>                          effect = "individual",
>                          model = "random", 
>                          lag = FALSE,
>                          spatial.error = "b",
>                          index = c("Group", "Year"))
>
> Error in .C64("aplsb1", SIGNATURE = c(SS$signature, SS$signature, 
> "double", : NAs in argument 7 and 'NAOK = FALSE' (dotCall64)

Please at least provide the output of traceback() after the error (it does 
not crash, it error-exits correctly). Also provide the output of 
sessionInfo() - versions of underlying packages may matter (I'm thinking 
of dotCall64 and packages it uses). .C64() is not called by splm functions 
directly.

Does the error persist on a fully updated system on the same OS?

What happens on a different platform if available?

Ideally, provide a fully reproducible minimal example using built-in data, 
or provide your data and the code used to create analysis.subset and 
listw.wgts, so that the running the code creating the error can be 
reproduced exactly. The error message reports NAs, are the data complete?

Roger

>
> Unlike the same specification but with fixed effects (with model = 
> "within" argument), the one above takes a long while to run until it 
> crashes. As far as I can tell, there are no memory issues.
>
> An alternative spatial random effects using spgm() command runs OK, as 
> do non-spatial fixed and random effects models using plm(). For the one 
> above I could use spgm(), but I would prefer to keep all my models in 
> the maximum likelihood estimation for consistency.
>
> Any clarifications as to the nature of the error are welcome. 
>
> Thank you.
>
>
> Denys Dukhovnov
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo using r-project.org
> https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fstat.ethz.ch%2Fmailman%2Flistinfo%2Fr-sig-geo&data=05%7C01%7CRoger.Bivand%40nhh.no%7Cd4a3c3a836f44c598f9608db0f6250a2%7C33a15b2f849941998d56f20b5aa91af2%7C0%7C0%7C638120687160554379%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Oa5KOUxlzjHxBzBfAd12W9WrTRVBynr1bYdJlST%2F17g%3D&reserved=0
>

-- 
Roger Bivand
Emeritus Professor
Department of Economics, Norwegian School of Economics,
Postboks 3490 Ytre Sandviken, 5045 Bergen, Norway.
e-mail: Roger.Bivand using nhh.no
https://orcid.org/0000-0003-2392-6140
https://scholar.google.no/citations?user=AWeghB0AAAAJ&hl=en