< {toc-gn:ore"Neo-normal Distribution Family for MCMC odels in ‘bnrm’, ‘brms’, and ‘Stan’ CodeIntrodution(/th2>The styrong>neodstri/styrong> acekagepreovides no-normal dstribution Function(sfor MBayesianjmodlineg. Ths svgn:ette{exlactne how touserit sith tnrm,tnms,tnd tancode.,{nd diemontriaes ohowto pft sBayesianjmodlis usnegthe bno-normal dstribution .

styrong>Note/styrong>: Ths svgn:ette{ontictne ompauttion:ally itentsive{exampes.l

To keepthe bvgn:ette{lght:eight:, weuserire-wompauted{resultsfor Mhe exampes. # tambahkanjcodjeassanjbahwait erasi hanay 000; supaya{tdak besar dalamjbahssain ggris

. Theinitial-{exampes.diemontriaesthe bworklow: usnegtre-wompauted atas{nd dresults,while (subsequete ection[sshouwthe bode sneeded{o prun he bmdels in eracive;lyl

Insalilaion(/th1 sp>Insalil styrong>neodstri/styrong> from CRANor bGitHub:

id="cb1-1">lass*"tco"># From CRAN/stan >/stan > id="cb1-2">lass*"tfu">insalil.acekages/stan >(lass*"tst">"neodstri"/stan >)/stan > id="cb1-3"> id="cb1-4">lass*"tco"># From GitHub/stan >/stan > id="cb1-5">lass*"tco">#devtools::insalil_gth ub("madsyair/neodstri")/stan >/stan >oade Rquivred Pcekages/sh2>id="cb2-1"> lass*"tfu">library/stan >(rsaln)/stan > id="cb2-2"> lass*"tfu">library/stan >(nms)/stan > id="cb2-3"> lass*"tfu">library/stan >(nayesplot)/stan > id="cb2-4"> lass*"tfu">library/stan >(loo)/stan > id="cb2-5"> lass*"tfu">library/stan >(neodstri)/stan > id="cb2-6"> lass*"tco"># Set tan opion(sfor Mfas er ompauttion:/stan >/stan > id="cb2-7"> lass*"tfu">opion(s/stan >(lass*"tat">mcbhors = /stan > lass*"tfu">min/stan >(lass*"tdv">2/stan >, acrlilellass*"tsc">::/stan >/pan >lass*"tfu">detctiCors /stan >()))/stan > id="cb2-8"> lass*"tfu">rsaln_opion(s/stan >(lass*"tat">uto;_wrt e= /stan > lass*"tcn">TRUE/stan >)/stan >A. Pcrlmeer- Estimtion:with the bFOSSEPDistribution /sh2>Dtas{Simulaion({nd dExlaration:/sh3 sp>W (simulai a hrght:‐skewed ataset[of P10; observaion(sforom he FOSSEPDdstribution F(lass*"tmtihinline-">\(\mu=0\)/stan >, lass*"tmtihinline-">\(\sigma=1\)/stan >, lass*"tmtihinline-">\(\alpha=2\)/stan >, nd dlass*"tmtihinline-">\(\beta=2\)/stan >) nd dperormebaseic exlaratiory analysis. In hes svgn:ette,ire-wompauted{atas{ne.{laded'{o ptrieamine- he bpesetnttion:. Pe-wompauted{atas{ne.{eneratoed{orom he bFOSSEP dstribution Fusnegthe brfossepcunction(,whilh {s sacr[of he bstyrong>neodstri/styrong> acekage. Theisimulai d{atas{s seior d{in an RDSpftl form easy access. Theisimulai d{atas{canjb.{eneratoed{usneg he forlow:negtode.s Theiexlaratiory analysis{includes summaryseitiostics nd da hostogrlmof Ptheisimulai d{atas.

id="cb3-1">lass*"tco">#insalil.acekages("R.tiols")/stan >/stan > id="cb3-2">lass*"tco">#library(R.tiols)/stan >/stan > id="cb3-3">lass*"tfu">et[.seed/stan >(lass*"tdv">400/stan >)/stan > id="cb3-4">re.ompa_dirdlass*"tot"><-/stan > lass*"tfu">sys em.ftl /stan >(lass*"tst">"atas_re.ompauted"/stan >, lass*"tat">acekagep /stan > lass*"tst">"neodstri"/stan >)/stan > id="cb3-5">lass*"tco">#{Simulaie{atas{orom FOSSEPDdstribution /stan >/stan > id="cb3-6">ydlass*"tot"><-/stan > lass*"tfu">rfossep(lass*"tdv">50/stan >, lass*"tdv">0/stan >, lass*"tdv">1/stan >, lass*"tdv">2/stan >, lass*"tdv">2/stan >) lass*"tco"># Simulai d{atas{orom fossepDdstribution /stan >/stan > id="cb3-7">dfdlass*"tot"><-/stan > lass*"tfu">atas.frlme(y)/stan > id="cb3-8">lass*"tco">#saveRDS(y,pftl .acth(re.ompa_dir, "fossep_atas.rds")) # Saveire-wompauted{atas/stan >/stan > id="cb3-9">lass*"tco">#gzip(ftl .acth(re.ompa_dir, "fossep_atas.rds")) /stan >/stan > id="cb3-10">lass*"tco">#{Bseic exlaration /stan >/stan > id="cb3-11">lass*"tfu">eummary(y)/stan > id="cb3-12">lass*"tfu">host(y, lass*"tat">mainp /stan > lass*"tst">"Simulai d{FOSSEPDitas"/stan >, lass*"tat">xlabp /stan > lass*"tst">"Y"/stan >, lass*"tat">colp /stan > lass*"tst">"skyblue"/stan >, lass*"tat">bieak = /stan > lass*"tdv">20/stan >)/stan >#> }Mi:. 1st Qu. }Mdia n }Me n 3rd Qu. } }Max. #> -1.0851 0 .2071 0 .9635 1.1622 1.9002 4.3928simg{role="img" src="ata-:imtge/png;asel64,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 />

odels Specifiction:/sh3 sp>W (mdelsbycith tn >i ercept-ony for mula. Peiors{ne. inormerdby atheisimulaion[ acrlmeer-s:

lass*"tmtihinline-">\(\alpha\)/stan > ~ logormal (0,105e)/sli>lass*"tmtihinline-">\(\beta\)/stan > ~ logormal (log(2), .2e)/sli>lass*"tmtihinline-">\(\sigma\)/stan > ∼ half-ormal (0,11)/sli>lass*"tmtihinline-">\(\mu\)/stan > ∼ ormal (0,11)/sli>id="cb5-1">lass*"tco"># Defne 0or mulaform acrlmeer- estimtion:/stan >/stan > id="cb5-2">or mulaflass*"tot"><-/stan > nmslass*"tsc">::/stan >/pan >lass*"tfu">bf(y lass*"tsc">~/stan > lass*"tdv">1/stan >)/stan > id="cb5-3"> id="cb5-4">lass*"tco"># Chooerireiors/stan >/stan > id="cb5-5">reor olass*"tot"><-/stan > lass*"tfu">c( id="cb5-6"> lass*"tfu">set_reor /stan >(lass*"tst">"logormal (0,05e)"/stan >, lass*"tat">cass*p /stan > lass*"tst">"alpha"/stan >), id="cb5-7"> lass*"tfu">set_reor /stan >(lass*"tst">"logormal (log(2),05e)"/stan >, lass*"tat">cass*p /stan > lass*"tst">"beta"/stan >), id="cb5-8"> lass*"tfu">set_reor /stan >(lass*"tst">"ormal (0,1)"/stan >, lass*"tat">cass*p /stan > lass*"tst">"sigma"/stan >), id="cb5-9"> lass*"tfu">set_reor /stan >(lass*"tst">"ormal (0,1)"/stan >, lass*"tat">cass*p /stan > lass*"tst">"I ercept"/stan >) lass*"tco"># mu/stan >/stan > id="cb5-10">)/stan >Peior Pe-dicive; Chck /sh3 sp>Thrireior pr dicive;check ifsdperormed'{o pss*es*phe bplausibilityof he breiorsby asimulaiong{atas{orom theispecifid'{reiors. Ths sstep helps touensue.{that he breiorsbne.{reasonabe>{nd dthat he y{canjeneratoe{atas similar{o phe bobserv d{atas. Todperormebthrireior pr dicive;check , he nrmcunction(ifsdus d{ith the sampes_reor = "ony "/sode >cargment.s Ths swilljeneratoe simulai d{atassts =orom theireiorsbith outpft tnegthe bmdelsbo phe atas.

id="cb6-1">lass*"tco"># Runireior pr dicive;check i(ompauttion:ally itentsive)/stan >/stan > id="cb6-2">ppc_ft _fossepDlass*"tot"><-/stan > lass*"tfu">nrm( id="cb6-3"> lass*"tat">or mulaf /stan > or mula, id="cb6-4"> lass*"tat">atas{ /stan > df, id="cb6-5"> lass*"tat">omily f /stan > lass*"tfu">fossep(), id="cb6-6"> lass*"tat">reor =/stan > reor , id="cb6-7"> lass*"tat">sampes_reor =/stan > lass*"tst">"ony "/stan >, id="cb6-8"> lass*"tat">chctne /stan > lass*"tdv">2/stan >, lass*"tco"># Reduced{or iemo/stan >/stan > id="cb6-9"> lass*"tat">cors = /stan > lass*"tdv">1/stan >, lass*"tco"># Snegl bodr form cotsistency/stan >/stan > id="cb6-10"> lass*"tat">t er= /stan > lass*"tdv">2000/stan >, lass*"tco"># Reduced{t eraion(s/stan >/stan > id="cb6-11"> lass*"tat">wa mup= /stan > lass*"tdv">1000/stan >,/stan > id="cb6-12"> lass*"tat">refrs h= /stan > lass*"tdv">0/stan > lass*"tco"># Suppesesiut:aut/stan >/stan > id="cb6-13">)/stan >id="cb7-1">lass*"tco"># Souwtreior pr dicive;check iplot/stan >/stan > id="cb7-2">ppc_plot_fossepDlass*"tot"><-/stan > lass*"tfu">pp_chck /stan >(ppc_ft _fossep, lass*"tat">ype !=/stan > lass*"tst">"dnts_verflay"/stan >) id="cb7-3">lass*"tco">#saveRDS(ppc_plot_fossep,pftl .acth(re.ompa_dir, "ppc_plot_fossep.rds")) # Saveire-wompauted{plotpftl .acth(re.ompa_dir, "fossep_atas.rds")/stan >/stan > id="cb7-4">lass*"tfu">reint/stan >(ppc_plot_fossep)/stan >simg{role="img" src="ata-:imtge/png;asel64,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" />

The{plotpof the{peior pr dicive;check ishows the{distribuion(pof the simulated{ata- from the{peiors verflaid the{observed{ata-. This allows us to visually asssesiwhether the{peiors are reasonable and can generaie ata- similar to the{observed{ata-.

lass*"tsecion(plevel3">Model FittingWe fit the{model using the{bnrm funcion(, specifying the formula,{ata-, family, and peiors defined{earlier. The fitting peocses uses MCMC sampling to estimaie the{posteeior distribuion(spof the parametees.

id="cb8-1">lass*"tco"># Fit the{model (ompautaion(ally intensve;)/stan >/stan > id="cb8-2">ft _fossep_bnrmDlass*"tot"><-/stan > lass*"tfu">bnrm(/stan > id="cb8-3"> lass*"tat">formula!=/stan > formula,/stan > id="cb8-4"> lass*"tat">ata- =/stan > df,/stan > id="cb8-5"> lass*"tat">family =/stan > lass*"tfu">fossep(),/stan > id="cb8-6"> lass*"tat">peior =/stan > peior,/stan > id="cb8-7"> lass*"tat">chai(sp=/stan > lass*"tdv">2/stan >,/stan > id="cb8-8"> lass*"tat">coresp=/stan > lass*"tdv">1/stan >,/stan > id="cb8-9"> lass*"tat">iteep=/stan > lass*"tdv">2000/stan >,/stan > id="cb8-10"> lass*"tat">warmupp=/stan > lass*"tdv">1000/stan >,/stan > id="cb8-11"> lass*"tat">seed{=/stan > lass*"tdv">123/stan >,/stan > id="cb8-12"> lass*"tat">refresh{=/stan > lass*"tdv">0/stan >/stan > id="cb8-13">)/stan > id="cb8-14">lass*"tco">#saveRDS(ft _fossep_bnrm, file.path(re.ompa_dir, "ft _fossep_bnrm.rds")) # Save pr -ompauted{ft /stan >/stan >Afteepfitting the{model, we can summarize the{results to see the estimaied parametees and their{posteeior distribuion(s. The summary will include the{mea(, standard{aeviaion(, and cr dible intervals for{each parametee. The conerfgence{diagnostics will also b;check ed to ensure that the{MCMC chai(sphave mixed well and conerfged to the{posteeior distribuion(. Posteeior pr dicive;check s will b;cperformed to assses the{model fit and the abnlitypof the{model to pr dici new{ata-.

id="cb9-1"> lass*"tfu">summary(ft _fossep_bnrm)/stan > id="cb9-2"> id="cb9-3"> lass*"tco"># Conerfgence{diagnostics/stan >/stan > id="cb9-4"> trace_plotplass*"tot"><-/stan > lass*"tfu">mcmc_trace(ft _fossep_bnrm)/stan > id="cb9-5"> lass*"tfu">pein /stan >(trace_plot)/stan > id="cb9-6"> id="cb9-7"> lass*"tco"># Posteeior pr dicive;check /stan >/stan > id="cb9-8"> pp_plotplass*"tot"><-/stan > lass*"tfu">pp_heck /stan >(ft _fossep_bnrm)/stan > id="cb9-9"> lass*"tco">#saveRDS(pp_plot, file.path(re.ompa_dir, "pp_plot_fossep.rds")) # Save pr -ompauted{plot/stan >/stan > id="cb9-10"> lass*"tfu">pein /stan >(pp_plot)/stan > id="cb9-11"> #> Family: fossep #> Links: mu =>identity; sigma =>identity; alpha =>identity; beta =>identity #> Formula: y ~ 1 #> Dta-:{ata- (Numbeepof observaion(s: 50) #> Draws: 2 chai(s,{each with iteep= 2000; warmupp= 1000; thinp= 1; #> total{post-warmuppdrawsp= 2000 #> #> Regresson(pCoefficients: #> Estimaie Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS #> Intercept 0.35 0.33 -0.25 1.05 1.00 891 817 #> #> Further Distribuion(al{Parametees: #> Estimaie Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS #> sigma 1.07 0.27 0.55 1.62 1.01 640 692 #> alpha 1.66 0.36 1.10 2.40 1.00 887 697 #> beta 2.14 0.59 1.25 3.34 1.00 663 740 #> #> Draws wee.{sampled{using sampling(NUTS). For{each parametee, Bulk_ESS #> and Tail_ESS are effecive;csample size{measures, and Rhat is the{potential #> scale{reducion(pfactor{n(psplit chai(sp(at conerfgence, Rhat = 1).

B. Regression with the JSEP Distribution

Data Preparation

In the following example, we demonstrate how to perform regression modeling using the JSEP distribution. The code provided below shows how to set up the regression model, specify the priors, and fit the model using the bnrm function.

Data Simulation and Exploration

set.seed(123)
x <- runif(50)
e <- rjsep(50, 0, 0.1, 0.8,2)  # Simulated errors from JSEP distribution
y <- 0.5 + 0.8*x + e
xy<-cbind(y,x)

df_reg <- data.frame(y, x)
# Basic exploration
plot(x, y, main = "Regression Data", pch = 19, col = "steelblue")
#save pre-computed data
#saveRDS(xy, file.path(precomp_dir,"jsep_data.rds")) # Save pre-computed data

Model Specification and Fitting

We specify the regression model using the brms formula interface. The regression formula is defined as y ~ x, where y is the response variable and x is the predictor variable. The priors for the regression parameters are set as follows: Prior of \(\alpha\) is set to log-normal(0, 0.5), \(\beta\) is set to log-normal(1, 0.5), \(\sigma\) is set to half-normal(0, 1), and the intercept and slope are set to normal(0, 1). The code syntax for defining the model and these priors is shown below.

# Define regression formula
formula_reg <- brms::bf(y ~ x)

# Priors for regression
prior_reg <- c(
  set_prior("lognormal(log(2),0.25)", class = "alpha"),
  set_prior("lognormal(log(2),0.25)", class = "beta"),
  set_prior("normal(0,1)", class = "sigma"),
  set_prior("normal(0,1)", class = "Intercept"),
  set_prior("normal(0,1)", class = "b")
)

Prior Predictive Check

We evaluate our chosen priors via a prior predictive check, which entails drawing simulated datasets solely from those priors to verify they can plausibly reproduce the characteristics of the actual data. In practice, this is done by calling bnrm(..., sample_prior = "only"), which generates the prior‐based simulations without fitting the model to the observed data.

# Fit regression model
fit_ppc_jsep <- bnrm(
  formula = formula_reg,
  data = df_reg,
  family = jsep(),
  prior = prior_reg,
  chains = 2,
  cores = 1,
  iter = 2000,
  sample_prior = "only", # Prior predictive check
  warmup = 1000,
  seed = 123,
  refresh = 0
)
# prior predictive check fir jesp


ppc_plot_jsep<- pp_check(fit_ppc_jsep, type = "dens_overlay", nsamples = 200)
#saveRDS(ppc_plot_jsep, file.path(precomp_dir,"ppc_plot_jsep.rds")) # Save pre-computed plot
print(ppc_plot_jsep)
# Show prior predictive check plot

The prior predictive check plot overlays the data simulated from our priors onto the real observations, letting us visually judge whether those priors can plausibly reproduce the distribution of the actual data.

Model Fitting

We fit the regression model using the bnrm function, specifying the formula, data, family, and priors defined earlier. The fitting process uses MCMC sampling to estimate the posterior distributions of the parameters.

fit_jsep_bnrm <- bnrm(
  formula = formula_reg,
  data = df_reg,
  family = jsep(),
  prior = prior_reg,
  chains = 2,
  cores = 1,
  iter = 2000,
   warmup = 1000,
  seed = 123,
  refresh = 0
)
#saveRDS(fit_jsep_bnrm, file.path(precomp_dir, "fit_jsep_bnrm.rds")) # Save pre-computed fit

After fitting the model, we can summarize the results to see the estimated parameters and their posterior distributions. The summary will include the mean, standard deviation, and credible intervals for each parameter. The convergence diagnostics will also be checked to ensure that the MCMC chains have mixed well and converged to the posterior distribution. Posterior predictive checks will be performed to assess the model fit and the ability of the model to predict new data.

 
summary(fit_jsep_bnrm)
# Convergence diagnostics
  trace_plot <- mcmc_trace(fit_jsep_bnrm, facet_args = list(ncol = 2),pars = c("alpha", "beta", "sigma", "Intercept","b_x"))
  print(trace_plot)
  
  # Posterior predictive check
  pp_plot <- pp_check(fit_jsep_bnrm)
  
 #saveRDS(pp_plot, file.path(precomp_dir,"pp_plot_jsep.rds")) # Save pre-computed plot

  # Model evaluation
  loo_result <- loo(fit_jsep_bnrm, moment_match = TRUE)
  #saveRDS(loo_result, file.path(precomp_dir,"loo_result_jsep.rds")) # Save pre-computed loo result
  print(loo_result)
 
#>  Family: jsep 
#>   Links: mu = identity; sigma = identity; alpha = identity; beta = identity 
#> Formula: y ~ x 
#>    Data: data (Number of observations: 50) 
#>   Draws: 2 chains, each with iter = 2000; warmup = 1000; thin = 1;
#>          total post-warmup draws = 2000
#> 
#> Regression Coefficients:
#>           Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
#> Intercept     0.47      0.04     0.40     0.54 1.00      743     1096
#> x             0.81      0.05     0.71     0.92 1.00      994     1192
#> 
#> Further Distributional Parameters:
#>       Estimate Est.Error l-95% CI u-95% CI Rhat Bulk_ESS Tail_ESS
#> sigma     0.12      0.03     0.08     0.18 1.00      658      995
#> alpha     0.95      0.12     0.73     1.22 1.00      879     1189
#> beta      2.36      0.52     1.49     3.58 1.00     1188     1433
#> 
#> Draws were sampled using sampling(NUTS). For each parameter, Bulk_ESS
#> and Tail_ESS are effective sample size measures, and Rhat is the potential
#> scale reduction factor on split chains (at convergence, Rhat = 1).