The order of the data cause different fitting result

Modeling
,
1

I changed the order of the data, sorted it by ID from smallest to largest, input it into the model after the same preprocessing, and got a completely different fitting result from before sorting. Why?

fit_m01_1 <- stan(file=paste0(getwd(),'/stan_model/m01_RW.stan'), data=data_list,
                             pars    = pars_m01_1,
                             chains  = 3,
                             iter    = 4000,
                             warmup  = 2000,
                             thin    = 1,
                             control = list(adapt_delta   = 0.95,
                                            stepsize      = 1,
                                            max_treedepth = 10))

微信图片_20240604175516
微信图片_20240604175516692×417 13.5 KB

微信图片_20240604175536
微信图片_20240604175536719×422 8.46 KB

2

Can you share the code for how you processed and sorted the data? It’s almost certainly occurring due to an error here.

3

Thank you very much.
The original data diagram is as shown in the figure. I sorted the data in Excel based on the subid in ascending order (starting from 1), and then performed the same processing in R. The code used in R is not convenient to show, but after sorting the original data 1, the fitting results are the same as data 2. However, the unsorted original data 1 does not fit the results.

1111
1111959×321 7.88 KB

4

I’m sorry, but I’d have to see the R code of reading in the data and sorting it before I can help diagnose the issue.

5

it‘s the code

图片1
图片1991×630 26.9 KB

图片2
图片21183×724 22.1 KB

6

Here is the code, can you provide some help?