Report bym in openbugs . mc-error , val97,5 pc , running quantiles,auto-corrolation

General
, ,
1

hi everyone, my friends, I’m having trouble reporting bym results in open bugs and I’m confused.
can you guide me in reporting MC_error and val97.5pc in below?

               mean	  sd	        MC_erro 	val2.5pc	median	val97.5pc	start	sample
theta[1]	0.6009	0.005858	4.171E-5	0.5897	0.6009	0.6127	20001	20000
theta[2]	0.3557	0.006764	4.513E-5	0.3425	0.3556	0.369	20001	20000
theta[3]	0.928	0.005863	4.006E-5	0.9164	0.928	0.9395	20001	20000
theta[4]	0.7481	0.00635	4.07E-5	0.7357	0.748	0.7605	20001	20000
theta[5]	0.2292	0.006414	4.522E-5	0.2168	0.2291	0.2419	20001	20000
theta[6]	0.2452	0.006796	4.915E-5	0.2323	0.245	0.2586	20001	20000
theta[7]	1.365	0.006228	4.449E-5	1.353	1.365	1.377	20001	20000
theta[8]	1.127	0.006132	4.116E-5	1.115	1.127	1.139	20001	20000
theta[9]	0.2759	0.006053	4.362E-5	0.264	0.2758	0.288	20001	20000
theta[10]	1.024	0.006904	4.114E-5	1.01	       1.023	1.037	20001	20000
theta[11]	0.4864	0.006848	5.004E-5	0.4731	0.4864	0.4999	20001	20000
theta[12]	0.5611	0.006956	5.157E-5	0.5476	0.561	0.5748	20001	20000
theta[13]	0.3082	0.006485	4.024E-5	0.2957	0.3082	0.3212	20001	20000
theta[14]	0.1332	0.006389	4.069E-5	0.1211	0.1331	0.146	20001	20000
theta[15]	0.6085	0.006085	4.196E-5	0.5966	0.6085	0.6206	20001	20000
theta[16]	0.5747	0.006599	4.02E-5	0.5618	0.5747	0.5877	20001	20000
theta[17]	0.1638	0.006091	 4.26E-5	0.152	0.1638	0.1758	20001	20000
theta[18]	1.28	        0.005913	4.028E-5	1.269	1.28	      1.292	        20001	20000
theta[19]	0.1978	0.00617	4.325E-5	0.1858	0.1977	0.2101	20001	20000
theta[20]	1.162	0.007063	5.048E-5	1.148	1.162	1.176	20001	20000
theta[21]	0.1471	0.006573	4.554E-5	0.1344	0.147	0.1603	20001	20000
theta[22]	0.3653	0.006177	4.391E-5	0.3533	0.3653	0.3774	20001	20000
theta[23]	0.4845	0.006884	4.906E-5	0.4711	0.4845	0.4981	20001	20000
theta[24]	0.3768	0.006266	4.26E-5	0.3646	0.3768	0.389	20001	20000
theta[25]	0.8934	0.005998	4.842E-5	0.8817	0.8934	0.9053	20001	20000
theta[26]	0.2425	0.005934	3.877E-5	0.231	0.2425	0.2543	20001	20000
theta[27]	0.173	0.006135	4.632E-5	0.1614	0.1729	0.1854	20001	20000
theta[28]	0.301	0.005738	4.131E-5	0.2899	0.301	0.3123	20001	20000
theta[29]	1.918	0.004515	3.036E-5	1.909	1.918	1.927	20001	20000
theta[30]	0.2376	0.006035	4.091E-5	0.2258	0.2376	0.2495	20001	20000
theta[31]	0.6009	0.005932	4.528E-5	0.5893	0.6009	0.6125	20001	20000

and running quantiles…
image

and auto-corrolation …

image

cheers.

2

maybe the answer is in the GeoBugs user’s guide? https://www.mrc-bsu.cam.ac.uk/wp-content/uploads/geobugs12manual.pdf

I’ve never used this model. for the paper on implementing the BYM2 model in Stan (which differs from the original BYM model, in ways discussed here: https://arxiv.org/pdf/1601.01180.pdf),
we did a comparison of a simpler model in Stan and winbugs which was just a poisson model with an ICAR component - not the same as the bym model.

if you want to use the BYM2 model, your choices are to use BRMS, INLA, or the models in this repo: GitHub - ConnorDonegan/Stan-IAR: Spatial intrinsic autoregressive models in Stan

trying to start with the GeoBUGS bym model as a baseline, while perhaps methodologically warrented, may well be impossible with your data.

1 Like
3

hi professor @mitzimorris, I managed to find the answer to my question. thanks for your help.
I run the bym2 model in stan according to your article and I also needed the bym results in openbugs .
now I want to compare the GIS models and change point models and spatial models in the estimation of disease changes.
is it possible to introduce the right link so that I can get more information about this?

4

what do you mean “right link”?

I’m far from an expert on the models you mention.

for doing model comparison, you might find examples and references in this paper helpful: [2011.01808] Bayesian Workflow

1 Like
5

i am still a beginner and it is very difficult to get results. thank you for your time.