brms

https://paulbuerkner.com/brms
Admin
brms is an R package for fitting Bayesian generalized (non-)linear multivariate multilevel models...
Tokens:193,650
Snippets:1,805
Trust Score:7.9
Update:3 months ago
Show doc for...
Tokens:
Raw
### Example of using threading in brm

Source: https://paulbuerkner.com/brms/reference/threading.html

This example demonstrates how to use the `threading` function with the `brm` function to fit a model with within-chain parallelization. Note that within-chain parallelization is experimental and may not always provide speedups. Setting `static = TRUE` in `threading` is recommended for exact reproducibility.

```r
if (FALSE) {
# this model just serves as an illustration
# threading may not actually speed things up here
fit <- brm(count ~ zAge + zBase * Trt + (1|patient),
           data = epilepsy, family = negbinomial(),
           chains = 1, threads = threading(2, grainsize = 100),
           backend = "cmdstanr")
summary(fit)
}
```

--------------------------------

### Install brms from GitHub

Source: https://paulbuerkner.com/brms/index.html

Installs the latest development version of the brms package from GitHub. Requires the 'remotes' package to be installed first.

```r
if (!requireNamespace("remotes")) {
  install.packages("remotes")
}
remotes::install_github("paul-buerkner/brms")
```

--------------------------------

### Posterior Summary Example

Source: https://paulbuerkner.com/brms/reference/posterior_summary.html

Demonstrates how to use the posterior_summary function with a brmsfit object. It also shows the recommended workflow using the 'posterior' package.

```R
if (FALSE) {
fit <- brm(time ~ age * sex, data = kidney)
posterior_summary(fit)

# recommended workflow using posterior
library(posterior)
draws <- as_draws_array(fit)
summarise_draws(draws, default_summary_measures())
}
```

--------------------------------

### Example Usage of loo_predictive_interval and loo_predict

Source: https://paulbuerkner.com/brms/reference/loo_predict.brmsfit.html

Demonstrates how to compute predictive intervals and variances using `loo_predictive_interval` and `loo_predict`. It shows how to optionally pre-compute PSIS weights for efficiency.

```R
## data from help("lm")
ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
d <- data.frame(
  weight = c(ctl, trt),
  group = gl(2, 10, 20, labels = c("Ctl", "Trt"))
)
fit <- brm(weight ~ group, data = d)
loo_predictive_interval(fit, prob = 0.8)

## optionally log-weights can be pre-computed and reused
psis <- loo::psis(-log_lik(fit), cores = 2)
loo_predictive_interval(fit, prob = 0.8, psis_object = psis)
loo_predict(fit, type = "var", psis_object = psis)

```

--------------------------------

### R2D2 Prior Setup

Source: https://paulbuerkner.com/brms/reference/R2D2.html

Use R2D2 to set up priors for population-level effects. The function itself does not evaluate arguments but assists in model setup.

```R
R2D2(mean_R2 = 0.5, prec_R2 = 2, cons_D2 = 0.5, autoscale = TRUE, main = FALSE)
```

--------------------------------

### Example: Fit a brms model with AR(2) correlation

Source: https://paulbuerkner.com/brms/reference/ar.html

This example demonstrates fitting a brms model using an AR(2) correlation structure on the 'x' variable from the LakeHuron dataset. Ensure the 'brms' package is installed and loaded.

```R
if (FALSE) {
data("LakeHuron")
LakeHuron <- as.data.frame(LakeHuron)
fit <- brm(x ~ ar(p = 2), data = LakeHuron)
summary(fit)
}
```

--------------------------------

### Example of using validate_prior with custom priors

Source: https://paulbuerkner.com/brms/reference/validate_prior.html

Demonstrates how to define custom priors using `set_prior` and `+` operator, and then validate them with `validate_prior` for a Poisson model.

```r
prior1 <- prior(normal(0,10), class = b) +
  prior(cauchy(0,2), class = sd)
validate_prior(prior1, count ~ zAge + zBase * Trt + (1|patient),
               data = epilepsy, family = poisson())
```

--------------------------------

### Example: Exact Cross-Validation for Problematic Observations

Source: https://paulbuerkner.com/brms/reference/reloo.brmsfit.html

This example demonstrates how to use the `reloo` function to compute exact cross-validation for a `brmsfit` object when the initial `loo` object indicates problematic Pareto k-estimates. It shows the typical workflow of fitting a model, computing LOO-CV, and then using `reloo` to refine the results for problematic observations.

```R
if (FALSE) {
fit1 <- brm(count ~ zAge + zBase * Trt + (1|patient),
            data = epilepsy, family = poisson())
# throws warning about some pareto k estimates being too high
(loo1 <- loo(fit1))
(reloo1 <- reloo(fit1, loo = loo1, chains = 1))
}
```

--------------------------------

### Example: Displaying All Smooth Terms - R

Source: https://paulbuerkner.com/brms/reference/conditional_smooths.brmsfit.html

Demonstrates how to display all smooth terms from a fitted brms model using conditional_smooths and plot the results.

```R
set.seed(0)
dat <- mgcv::gamSim(1, n = 200, scale = 2)
fit <- brm(y ~ s(x0) + s(x1) + s(x2) + s(x3), data = dat)
# show all smooth terms
plot(conditional_smooths(fit), rug = TRUE, ask = FALSE)
```

--------------------------------

### Example: Basic Usage and with New Data

Source: https://paulbuerkner.com/brms/reference/bayes_R2.brmsfit.html

Demonstrates how to fit a `brms` model and then compute its Bayesian R-squared. It also shows how to calculate R-squared using a new dataset.

```R
if (FALSE) {
fit <- brm(mpg ~ wt + cyl, data = mtcars)
summary(fit)
bayes_R2(fit)

# compute R2 with new data
nd <- data.frame(mpg = c(10, 20, 30), wt = c(4, 3, 2), cyl = c(8, 6, 4))
bayes_R2(fit, newdata = nd)
}
```

--------------------------------

### Running loo_subsample on a brmsfit model

Source: https://paulbuerkner.com/brms/reference/loo_subsample.brmsfit.html

Example of how to run loo_subsample on a brmsfit object. This is useful for models with population-level effects only.

```R
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler)
(loo1 <- loo_subsample(fit1))
```

--------------------------------

### Example: Disable Group-Level Coefficients

Source: https://paulbuerkner.com/brms/reference/save_pars.html

This example demonstrates how to prevent the saving of group-level coefficients to save memory. It shows the usage within the `brm` function and how to inspect the saved variables.

```R
if (FALSE) {
# don't store group-level coefficients
fit <- brm(count ~ zAge + zBase * Trt + (1|patient),
           data = epilepsy, family = poisson(),
           save_pars = save_pars(group = FALSE))
variables(fit)
}
```

--------------------------------

### Example: UNSTR correlation for visits within patients

Source: https://paulbuerkner.com/brms/reference/unstr.html

This example demonstrates adding an unstructured correlation matrix for visits within the same patient using `unstr()` in a `brm()` model formula.

```R
if (FALSE) {
# add an unstructured correlation matrix for visits within the same patient
fit <- brm(count ~ Trt + unstr(visit, patient), data = epilepsy)
summary(fit)
}
```

--------------------------------

### Example Usage of VarCorr for brmsfit

Source: https://paulbuerkner.com/brms/reference/VarCorr.brmsfit.html

Demonstrates how to use the VarCorr function to extract variance and correlation components from a fitted brmsfit model. This example requires a pre-fitted model object.

```R
if (FALSE) {
fit <- brm(count ~ zAge + zBase * Trt + (1+Trt|visit),
           data = epilepsy, family = gaussian(), chains = 2)
VarCorr(fit)
}
```

--------------------------------

### Example: Compute Posterior Intervals

Source: https://paulbuerkner.com/brms/reference/posterior_interval.brmsfit.html

Demonstrates how to fit a brms model and then compute its posterior intervals. This is useful for understanding the uncertainty around model parameters.

```R
if (FALSE) {
fit <- brm(count ~ zAge + zBase * Trt,
           data = epilepsy, family = negbinomial())
posterior_interval(fit)
}
```

--------------------------------

### Install brms from CRAN

Source: https://paulbuerkner.com/brms/index.html

Installs the latest stable release of the brms package from CRAN. This is the recommended method for most users.

```r
install.packages("brms")
```

--------------------------------

### Example: Extracting Diagnostic Quantities from a brmsfit Object

Source: https://paulbuerkner.com/brms/reference/diagnostic-quantities.html

Demonstrates how to fit a brms model and then extract and inspect diagnostic quantities such as log-posterior, NUTS parameters, R-hat, and neff_ratio.

```R
if (FALSE) {
fit <- brm(time ~ age * sex, data = kidney)

lp <- log_posterior(fit)
head(lp)

np <- nuts_params(fit)
str(np)
# extract the number of divergence transitions
sum(subset(np, Parameter == "divergent__")$Value)

head(rhat(fit))
head(neff_ratio(fit))
}
```

--------------------------------

### Compute Posterior Model Probabilities

Source: https://paulbuerkner.com/brms/reference/post_prob.brmsfit.html

This example demonstrates computing posterior model probabilities for two brmsfit models. Ensure `save_all_pars = TRUE` was set when fitting the models.

```R
if (FALSE) {
# model with the treatment effect
fit1 <- brm(
  count ~ zAge + zBase + Trt,
  data = epilepsy, family = negbinomial(),
  prior = prior(normal(0, 1), class = b),
  save_all_pars = TRUE
)
summary(fit1)

# model without the treatent effect
fit2 <- brm(
  count ~ zAge + zBase,
  data = epilepsy, family = negbinomial(),
  prior = prior(normal(0, 1), class = b),
  save_all_pars = TRUE
)
summary(fit2)

# compute the posterior model probabilities
post_prob(fit1, fit2)

# specify prior model probabilities
post_prob(fit1, fit2, prior_prob = c(0.8, 0.2))
}
```

--------------------------------

### Compute and Inspect Expected Predictions

Source: https://paulbuerkner.com/brms/reference/fitted.brmsfit.html

This example demonstrates fitting a brms model, computing its expected fitted values using the `fitted` method, and inspecting the first few results.

```R
## fit a model
fit <- brm(rating ~ treat + period + carry + (1|subject),
           data = inhaler)

## compute expected predictions
fitted_values <- fitted(fit)
head(fitted_values)
```

--------------------------------

### K-Fold Cross-Validation and Prediction Example

Source: https://paulbuerkner.com/brms/reference/kfold_predict.html

This example demonstrates how to perform K-fold cross-validation, save the fitted models, define a custom loss function (RMSE), and then use kfold_predict to obtain predictions for evaluation.

```R
if (FALSE) {
fit <- brm(count ~ zBase * Trt + (1|patient),
           data = epilepsy, family = poisson())

# perform k-fold cross validation
(kf <- kfold(fit, save_fits = TRUE, chains = 1))

# define a loss function
rmse <- function(y, yrep) {
  yrep_mean <- colMeans(yrep)
  sqrt(mean((yrep_mean - y)^2))
}

# predict responses and evaluate the loss
kfp <- kfold_predict(kf)
rmse(y = kfp$y, yrep = kfp$yrep)
}
```

--------------------------------

### Example: Compare WAIC of Two Models

Source: https://paulbuerkner.com/brms/reference/compare_ic.html

Demonstrates how to compare the WAIC of two `brmsfit` models. Ensure models are fitted and WAIC is computed before comparison. This functionality is superseded by `loo_compare`.

```R
if (FALSE) {
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler)
waic1 <- waic(fit1)

# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler)
waic2 <- waic(fit2)

# compare both models
compare_ic(waic1, waic2)
}
```

--------------------------------

### Compute and compare WAIC for brmsfit models

Source: https://paulbuerkner.com/brms/reference/waic.brmsfit.html

Example demonstrating how to fit two brms models and then compute and compare their WAIC values. This is useful for model selection.

```R
if (FALSE) {
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler)
(waic1 <- waic(fit1))

# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler)
(waic2 <- waic(fit2))

# compare both models
loo_compare(waic1, waic2)
}
```

--------------------------------

### Compute Model Weights for brmsfit Objects

Source: https://paulbuerkner.com/brms/reference/loo_model_weights.brmsfit.html

Example demonstrating how to compute model weights for two brmsfit objects using loo_model_weights. Ensure brmsfit objects are created before calling this function.

```R
if (FALSE) {
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler, family = "gaussian")
# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler, family = "gaussian")
loo_model_weights(fit1, fit2)
}
```

--------------------------------

### Vectorized Prior Definitions

Source: https://paulbuerkner.com/brms/reference/set_prior.html

This example shows how to define priors in a vectorized manner using `set_prior()`, which is particularly useful for categorical or multivariate models. The output displays the structure of the defined priors.

```R
set_prior("normal(0, 2)", dpar = c("muX", "muY", "muZ"))
```

```text
        prior class coef group resp dpar nlpar   lb   ub source
 normal(0, 2)     b                  muX       <NA> <NA>   user
 normal(0, 2)     b                  muY       <NA> <NA>   user
 normal(0, 2)     b                  muZ       <NA> <NA>   user
```

--------------------------------

### Fit mixture model with different predictors per component

Source: https://paulbuerkner.com/brms/reference/mixture.html

This example shows how to specify different predictor variables for each component of the mixture model using `brmsformula`.

```R
## use different predictors for the components
fit2 <- brm(bf(y ~ 1, mu1 ~ x, mu2 ~ z), dat, family = mix,
            prior = prior, chains = 2)
summary(fit2)
```

--------------------------------

### Example Usage of PSIS with brmsfit

Source: https://paulbuerkner.com/brms/reference/psis.brmsfit.html

Demonstrates how to fit a brms model and then apply the psis function to it for model evaluation. This is a common workflow for Bayesian model assessment.

```R
if (FALSE) {
fit <- brm(rating ~ treat + period + carry, data = inhaler)
psis(fit)
}
```

--------------------------------

### Example: Plotting Two-Dimensional Smooth Terms - R

Source: https://paulbuerkner.com/brms/reference/conditional_smooths.brmsfit.html

Illustrates fitting a model with a two-dimensional smooth term and plotting it using contour or raster styles.

```R
# fit and plot a two-dimensional smooth term
fit2 <- brm(y ~ t2(x0, x2), data = dat)
ms <- conditional_smooths(fit2)
plot(ms, stype = "contour")
plot(ms, stype = "raster")
```

--------------------------------

### Update Model with Future Package for Parallelization

Source: https://paulbuerkner.com/brms/reference/brm.html

Demonstrates using the 'future' package to set up parallel processing for updating a brms model. Requires the 'future' package to be installed.

```R
library(future)
plan(multisession, workers = 4)
fit7 <- update(fit7, future = TRUE)
```

--------------------------------

### Example: Displaying a Specific Smooth Term - R

Source: https://paulbuerkner.com/brms/reference/conditional_smooths.brmsfit.html

Shows how to display a specific smooth term, 's(x2)', from a fitted brms model using the 'smooths' argument.

```R
# show only the smooth term s(x2)
plot(conditional_smooths(fit, smooths = "s(x2)"), ask = FALSE)
```

--------------------------------

### Running loo_subsample on a brmsfit model with varying intercepts

Source: https://paulbuerkner.com/brms/reference/loo_subsample.brmsfit.html

Example of how to run loo_subsample on a brmsfit object that includes a varying intercept for subjects. This demonstrates handling more complex model structures.

```R
# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler)
(loo2 <- loo_subsample(fit2))
```

--------------------------------

### Example of using ma() in brms model fitting

Source: https://paulbuerkner.com/brms/reference/ma.html

This example demonstrates how to fit a brms model using a second-order Moving Average (MA(2)) correlation structure. Ensure the 'brms' package is installed and loaded, and the 'LakeHuron' dataset is available.

```R
if (FALSE) {
data("LakeHuron")
LakeHuron <- as.data.frame(LakeHuron)
fit <- brm(x ~ ma(p = 2), data = LakeHuron)
summary(fit)
}
```

--------------------------------

### Example: Applying loo_moment_match to a brmsfit object

Source: https://paulbuerkner.com/brms/reference/loo_moment_match.brmsfit.html

Demonstrates fitting a brms model with `save_pars = save_pars(all = TRUE)`, calculating LOO, and then applying moment matching to improve LOO-CV estimates. This is useful when standard LOO warnings about high Pareto k estimates are encountered.

```R
if (FALSE) {
fit1 <- brm(count ~ zAge + zBase * Trt + (1|patient),
            data = epilepsy, family = poisson(),
            save_pars = save_pars(all = TRUE))

# throws warning about some pareto k estimates being too high
(loo1 <- loo(fit1))
(mmloo1 <- loo_moment_match(fit1, loo = loo1))
}
```

--------------------------------

### Compute Predictive Intervals

Source: https://paulbuerkner.com/brms/reference/predictive_interval.brmsfit.html

Example of computing predictive intervals from a fitted brmsfit object. Ensure the brms package is installed and a model is fitted before running.

```R
if (FALSE) {
fit <- brm(count ~ zBase, data = epilepsy, family = poisson())
predictive_interval(fit)
}
```

--------------------------------

### Multi-membership model with level-specific covariate values

Source: https://paulbuerkner.com/brms/reference/mmc.html

This example demonstrates how to simulate data and fit a multi-membership model using brms, where covariates vary by group level using the mmc function. Ensure the brms package is installed and loaded.

```R
if (FALSE) {
# simulate some data
dat <- data.frame(
  y = rnorm(100), x1 = rnorm(100), x2 = rnorm(100),
  g1 = sample(1:10, 100, TRUE), g2 = sample(1:10, 100, TRUE)
)

# multi-membership model with level specific covariate values
dat$xc <- (dat$x1 + dat$x2) / 2
fit <- brm(y ~ xc + (1 + mmc(x1, x2) | mm(g1, g2)), data = dat)
summary(fit)
}
```

--------------------------------

### Example: Integrating CmdStanR fit into brms

Source: https://paulbuerkner.com/brms/reference/read_csv_as_stanfit.html

Demonstrates how to fit a model using CmdStanR and then use read_csv_as_stanfit to integrate the CmdStanR fit into a brmsfit object. This is useful for advanced workflows where direct CmdStanR fitting is preferred.

```R
if (FALSE) {
# fit a model manually via cmdstanr
scode <- stancode(count ~ Trt, data = epilepsy)
sdata <- standata(count ~ Trt, data = epilepsy)
mod <- cmdstanr::cmdstan_model(cmdstanr::write_stan_file(scode))
stanfit <- mod$sample(data = sdata)

# feed the Stan model back into brms
fit <- brm(count ~ Trt, data = epilepsy, empty = TRUE, backend = 'cmdstanr')
fit$fit <- read_csv_as_stanfit(stanfit$output_files(), model = mod)
fit <- rename_pars(fit)
summary(fit)
}
```

--------------------------------

### Performing and Comparing LOO Cross-Validation with brmsfit Models

Source: https://paulbuerkner.com/brms/reference/loo.brmsfit.html

This example demonstrates fitting two `brmsfit` models and then performing approximate leave-one-out cross-validation on each. Finally, it shows how to compare the models using `loo_compare`.

```R
if (FALSE) {
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler)
(loo1 <- loo(fit1))

# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler)
(loo2 <- loo(fit2))

# compare both models
loo_compare(loo1, loo2)
}
```

--------------------------------

### Basic emmeans analysis with brmsfit

Source: https://paulbuerkner.com/brms/reference/emmeans-brms-helpers.html

Demonstrates a typical workflow using brmsfit with emmeans. First, a brms model is fitted, then `ref_grid` is used to create an object compatible with emmeans, and finally, `emmeans` is applied to calculate estimated marginal means for a specific factor.

```R
if (FALSE) {
fit1 <- brm(time | cens(censored) ~ age * sex + disease + (1|patient),
            data = kidney, family = lognormal())
summary(fit1)

# summarize via 'emmeans'
library(emmeans)
rg <- ref_grid(fit1)
em <- emmeans(rg, "disease")
summary(em, point.est = mean)
}
```

--------------------------------

### Simulate Data and Fit Normal Mixture Model

Source: https://paulbuerkner.com/brms/reference/pp_mixture.brmsfit.html

This example demonstrates how to simulate data for a mixture model and fit a simple two-component normal mixture model using `brms`. It includes setting up priors for the model parameters.

```R
## simulate some data
set.seed(1234)
dat <- data.frame(
  y = c(rnorm(100), rnorm(50, 2)),
  x = rnorm(150)
)
## fit a simple normal mixture model
mix <- mixture(gaussian, nmix = 2)
prior <- c(
  prior(normal(0, 5), Intercept, nlpar = mu1),
  prior(normal(0, 5), Intercept, nlpar = mu2),
  prior(dirichlet(2, 2), theta)
)
fit1 <- brm(bf(y ~ x), dat, family = mix,
            prior = prior, chains = 2, init = 0)
summary(fit1)
```

--------------------------------

### Example Usage of cor_fixed

Source: https://paulbuerkner.com/brms/reference/cor_fixed.html

This example demonstrates how to use `cor_fixed` within a `brm` model to incorporate a predefined covariance matrix `V` for the response variable `y`. This is useful for meta-analysis scenarios.

```R
if (FALSE) {
dat <- data.frame(y = rnorm(3))
V <- cbind(c(0.5, 0.3, 0.2), c(0.3, 1, 0.1), c(0.2, 0.1, 0.2))
fit <- brm(y~1, data = dat, autocor = cor_fixed(V))
}
```

--------------------------------

### Using Alias Functions for Setting Priors

Source: https://paulbuerkner.com/brms/reference/set_prior.html

Demonstrates the usage of `prior()`, `prior_()`, and `prior_string()` as aliases for `set_prior` to define priors with different argument specification styles. It verifies that all three functions produce identical results.

```R
## use alias functions
(prior1 <- prior(cauchy(0, 1), class = sd))
#> sd ~ cauchy(0, 1)
(prior2 <- prior_(~cauchy(0, 1), class = ~sd))
#> sd ~ cauchy(0, 1)
(prior3 <- prior_string("cauchy(0, 1)", class = "sd"))
#> sd ~ cauchy(0, 1)
identical(prior1, prior2)
#> [1] TRUE
identical(prior1, prior3)
#> [1] TRUE
```

--------------------------------

### Custom Posterior Prediction Function Example

Source: https://paulbuerkner.com/brms/reference/get_dpar.html

Example of a custom posterior prediction function that uses `get_dpar` to retrieve draws for 'mu' and 'mypar' distributional parameters. This is useful when defining custom families in brms.

```R
posterior_predict_my_dist <- function(i, prep, ...) {
  mu <- brms::get_dpar(prep, "mu", i = i)
  mypar <- brms::get_dpar(prep, "mypar", i = i)
  my_rng(mu, mypar)
}
```

--------------------------------

### Model Averaging Example with brmsfit Objects

Source: https://paulbuerkner.com/brms/reference/pp_average.brmsfit.html

Demonstrates how to compute model-averaged predicted values using `pp_average` with two fitted brmsfit models and custom newdata. Requires prior fitting of brmsfit models.

```R
if (FALSE) {
# model with 'treat' as predictor
fit1 <- brm(rating ~ treat + period + carry, data = inhaler)
summary(fit1)

# model without 'treat' as predictor
fit2 <- brm(rating ~ period + carry, data = inhaler)
summary(fit2)

# compute model-averaged predicted values
(df <- unique(inhaler[, c("treat", "period", "carry")]))
pp_average(fit1, fit2, newdata = df)

# compute model-averaged fitted values
pp_average(fit1, fit2, method = "fitted", newdata = df)
}
```

--------------------------------

### Example AR(p) Correlation Structure with Grouping

Source: https://paulbuerkner.com/brms/reference/cor_ar.html

Example of constructing an AR(p) correlation structure with a specified time covariate ('visit') and a grouping factor ('patient'), setting the AR order to 2.

```R
cor_ar(~visit|patient, p = 2)
#> arma(~visit | patient, 2, 0)
```

--------------------------------

### Compute Model Weights using Stacking

Source: https://paulbuerkner.com/brms/reference/model_weights.brmsfit.html

Demonstrates how to compute model weights using the stacking method. This involves fitting two `brmsfit` models and then calling `model_weights` without specifying the `weights` argument, defaulting to 'stacking'.

```R
if (FALSE) {
# model with 'treat' as predictor
fit1 <- brm(rating ~ treat + period + carry, data = inhaler)
summary(fit1)

# model without 'treat' as predictor
fit2 <- brm(rating ~ period + carry, data = inhaler)
summary(fit2)

# obtain Akaike weights based on the WAIC
model_weights(fit1, fit2, weights = "waic")
}
```

--------------------------------

### Load nhanes Data

Source: https://paulbuerkner.com/brms/articles/brms_missings.html

Loads the 'nhanes' dataset from the 'mice' package for use in examples. Displays the first few rows of the dataset.

```r
data("nhanes", package = "mice")
head(nhanes)
```

--------------------------------

### Create an Empty Prior Object

Source: https://paulbuerkner.com/brms/reference/set_prior.html

The `empty_prior` function creates an empty prior object, which can be useful as a starting point or placeholder.

```R
empty_prior()
```

--------------------------------

### Get Variables from brmsfit Object

Source: https://paulbuerkner.com/brms/reference/draws-index-brms.html

Extracts the names of the variables from a brmsfit object. This method is part of the 'brms' package for R.

```R
variables(x, ...)
```

--------------------------------

### Set Student-t Prior for Specific Coefficient

Source: https://paulbuerkner.com/brms/reference/set_prior.html

Example of setting a unit student-t prior with 10 degrees of freedom for the coefficient 'x2'.

```R
set_prior("student_t(10, 0, 1)", class = "b", coef = "x2")
```

--------------------------------

### Set Normal Prior for Specific Coefficient

Source: https://paulbuerkner.com/brms/reference/set_prior.html

Example of setting a normal prior with mean 0 and standard deviation 5 for the coefficient 'x1'.

```R
set_prior("normal(0,5)", class = "b", coef = "x1")
```

--------------------------------

### Example Usage of posterior_table

Source: https://paulbuerkner.com/brms/reference/posterior_table.html

Demonstrates how to use `posterior_table` with a fitted brms model and its predictions. This is typically used for summarizing predictions of ordinal models.

```r
if (FALSE) {
fit <- brm(rating ~ period + carry + treat,
           data = inhaler, family = cumulative())
pr <- predict(fit, summary = FALSE)
posterior_table(pr)
}
```

--------------------------------

### Comparing Bayesian Models with WAIC

Source: https://paulbuerkner.com/brms/reference/loo_compare.brmsfit.html

Example demonstrating how to fit two Bayesian models using `brms` and compare them using the WAIC criterion with `loo_compare`. Ensure models have precomputed criteria using `add_criterion`.

```R
if (FALSE) {
# model with population-level effects only
fit1 <- brm(rating ~ treat + period + carry,
            data = inhaler)
fit1 <- add_criterion(fit1, "waic")

# model with an additional varying intercept for subjects
fit2 <- brm(rating ~ treat + period + carry + (1|subject),
            data = inhaler)
fit2 <- add_criterion(fit2, "waic")

# compare both models
loo_compare(fit1, fit2, criterion = "waic")
}
```

--------------------------------

### launch_shinystan.brmsfit

Source: https://paulbuerkner.com/brms/reference/launch_shinystan.brmsfit.html

Launches the ShinyStan application for a brmsfit object, enabling interactive model exploration.

```APIDOC
## Function: launch_shinystan.brmsfit

### Description
Provides an interface to shinystan for models fitted with brms.

### Signature
```R
launch_shinystan.brmsfit(object, rstudio = getOption("shinystan.rstudio"), ...)
```

### Arguments

*   **object** (brmsfit) - A fitted model object typically of class `brmsfit`.
*   **rstudio** (logical) - Only relevant for RStudio users. The default (`rstudio=FALSE`) is to launch the app in the default web browser rather than RStudio's pop-up Viewer. Users can change the default to `TRUE` by setting the global option `options(shinystan.rstudio = TRUE)`.
*   **...** - Optional arguments to pass to `runApp`.

### Value
An S4 shinystan object.

### See also
`launch_shinystan`

### Examples
```R
if (FALSE) {
fit <- brm(rating ~ treat + period + carry + (1|subject), 
           data = inhaler, family = "gaussian")
launch_shinystan(fit)
}
```
```

--------------------------------

### Plot Expected Predictions Against Actual Response

Source: https://paulbuerkner.com/brms/reference/fitted.brmsfit.html

This example visualizes the relationship between the actual response variable and the computed expected predictions from a brmsfit model.

```R
## plot expected predictions against actual response
dat <- as.data.frame(cbind(Y = standata(fit)$Y, fitted_values))
ggplot(dat) + geom_point(aes(x = Estimate, y = Y))
```

--------------------------------

### Get Number of Variables from brmsfit Object

Source: https://paulbuerkner.com/brms/reference/draws-index-brms.html

Returns the total number of variables in a brmsfit object. This method is part of the 'brms' package for R.

```R
nvariables(x, ...)
```

--------------------------------

### Create a Student's t family object with log link

Source: https://paulbuerkner.com/brms/reference/brmsfamily.html

Demonstrates how to create a family object for the Student's t distribution with a log link function using both the direct family function and the generic `brmsfamily` function. Note that these two methods do not produce identical objects.

```R
# create a family object
(fam1 <- student("log"))
#
#
# Family: student 
# Link function: log 
#

```

```R
# alternatively use the brmsfamily function
(fam2 <- brmsfamily("student", "log"))
#
#
# Family: student 
# Link function: log 
#

```

```R
# both leads to the same object
identical(fam1, fam2)
# [1] FALSE

```

--------------------------------

### Fit a Spatial CAR Model in brms

Source: https://paulbuerkner.com/brms/reference/car.html

This example demonstrates how to generate spatial data, set up adjacency matrices, and fit a binomial CAR model using brms. Ensure necessary packages for data generation and model fitting are loaded.

```R
if (FALSE) {
# generate some spatial data
east <- north <- 1:10
Grid <- expand.grid(east, north)
K <- nrow(Grid)

# set up distance and neighbourhood matrices
distance <- as.matrix(dist(Grid))
W <- array(0, c(K, K))
W[distance == 1] <- 1

# generate the covariates and response data
x1 <- rnorm(K)
x2 <- rnorm(K)
theta <- rnorm(K, sd = 0.05)
phi <- rmulti_normal(
  1, mu = rep(0, K), Sigma = 0.4 * exp(-0.1 * distance)
)
eta <- x1 + x2 + phi
prob <- exp(eta) / (1 + exp(eta))
size <- rep(50, K)
y <- rbinom(n = K, size = size, prob = prob)
dat <- data.frame(y, size, x1, x2)

# fit a CAR model
fit <- brm(y | trials(size) ~ x1 + x2 + car(W), 
           data = dat, data2 = list(W = W),
           family = binomial())
summary(fit)
}
```

--------------------------------

### Fit Probit Regression with Binomial Family

Source: https://paulbuerkner.com/brms/reference/brm.html

Example of fitting a probit regression model for binomial data. Requires data with trials and success counts.

```R
ntrials <- sample(1:10, 100, TRUE)
success <- rbinom(100, size = ntrials, prob = 0.4)
x <- rnorm(100)
data4 <- data.frame(ntrials, success, x)
fit4 <- brm(success | trials(ntrials) ~ x, data = data4,
            family = binomial("probit"))
summary(fit4)
```

--------------------------------

### Example: Extracting posterior draws from a brmsfit object

Source: https://paulbuerkner.com/brms/reference/draws-brms.html

Demonstrates how to convert a fitted brms model object into an array format for posterior analysis and how to extract specific variables using regular expressions.

```R
if (FALSE) {
fit <- brm(count ~ zAge + zBase * Trt + (1|patient),
           data = epilepsy, family = poisson())

# extract posterior draws in an array format
(draws_fit <- as_draws_array(fit))
posterior::summarize_draws(draws_fit)

# extract only certain variables
as_draws_array(fit, variable = "r_patient")
as_draws_array(fit, variable = "^b_", regex = TRUE)

# extract posterior draws in a random variables format
as_draws_rvars(fit)
}
```

--------------------------------

### Fit a simple GP model

Source: https://paulbuerkner.com/brms/reference/gp.html

This snippet demonstrates fitting a basic Gaussian process model using the 'gp()' function with a single predictor. Ensure data is simulated using 'mgcv::gamSim'.

```R
if (FALSE) {
# simulate data using the mgcv package
dat <- mgcv::gamSim(1, n = 30, scale = 2)

# fit a simple GP model
fit1 <- brm(y ~ gp(x2), dat, chains = 2)
summary(fit1)
me1 <- conditional_effects(fit1, ndraws = 200, spaghetti = TRUE)
plot(me1, ask = FALSE, points = TRUE)
}
```

--------------------------------

### Plotting Posterior Histograms

Source: https://paulbuerkner.com/brms/reference/mcmc_plot.brmsfit.html

Illustrates how to visualize the posterior distributions as histograms.

```R
# show histograms of the posterior distributions
mcmc_plot(model, type = "hist")
```

--------------------------------

### Get Number of Iterations from brmsfit Object

Source: https://paulbuerkner.com/brms/reference/draws-index-brms.html

Retrieves the number of iterations performed for each chain in a brmsfit object. This method is part of the 'brms' package for R.

```R
niterations(x)
```

--------------------------------

### Specifying R2D2 Prior Across Multiple Parameter Classes

Source: https://paulbuerkner.com/brms/reference/R2D2.html

Demonstrates how to apply the R2D2 prior across different parameter classes like 'b' (regression coefficients) and 'sd' (standard deviations). The 'main = TRUE' argument is crucial for specifying joint parameters like 'cons_D2'.

```R
set_prior(R2D2(mean_R2 = 0.8, prec_R2 = 10, main = TRUE), class = "b") +
  set_prior(R2D2(), class = "sd")
```

--------------------------------

### horseshoe()

Source: https://paulbuerkner.com/brms/reference/horseshoe.html

Sets up regularized horseshoe priors and related hierarchical shrinkage priors for population-level effects in brms. This function is used for model setup and does not evaluate its arguments.

```APIDOC
## horseshoe()

### Description
Sets up regularized horseshoe priors and related hierarchical shrinkage priors for population-level effects in brms. This function is used for model setup and does not evaluate its arguments.

### Function Signature
horseshoe(
  df = 1,
  scale_global = 1,
  df_global = 1,
  scale_slab = 2,
  df_slab = 4,
  par_ratio = NULL,
  autoscale = TRUE,
  main = FALSE
)

### Arguments

* **df** (numeric) - Degrees of freedom of student-t prior of the local shrinkage parameters. Defaults to `1`.

* **scale_global** (numeric) - Scale of the student-t prior of the global shrinkage parameter. Defaults to `1`. In linear models, `scale_global` will internally be multiplied by the residual standard deviation parameter `sigma`.

* **df_global** (numeric) - Degrees of freedom of student-t prior of the global shrinkage parameter. Defaults to `1`. If `df_global` is greater than `1`, the shape of the prior will no longer resemble a horseshoe and it may be more appropriately called an hierarchical shrinkage prior in this case.

* **scale_slab** (numeric) - Scale of the Student-t slab. Defaults to `2`. The original unregularized horseshoe prior is obtained by setting `scale_slab` to infinite, which can be approximated in practice by setting it to a very large real value.

* **df_slab** (numeric) - Degrees of freedom of the student-t slab. Defaults to `4`.

* **par_ratio** (numeric or NULL) - Ratio of the expected number of non-zero coefficients to the expected number of zero coefficients. If specified, `scale_global` is ignored and internally computed as `par_ratio / sqrt(N)`, where `N` is the total number of observations in the data.

* **autoscale** (logical) - Indicating whether the horseshoe prior should be scaled using the residual standard deviation `sigma` if possible and sensible (defaults to `TRUE`). Autoscaling is not applied for distributional parameters or when the model does not contain the parameter `sigma`.

* **main** (logical) - Not described in the source text.

```

--------------------------------

### Probit Regression with Binomial Family Example

Source: https://paulbuerkner.com/brms/reference/addition-terms.html

Illustrates fitting a probit regression model using the binomial family in brms. It shows how to specify the number of trials for each observation using resp_trials (implicitly via the trials() argument).

```R
## Probit regression using the binomial family
n <- sample(1:10, 100, TRUE)  # number of trials
success <- rbinom(100, size = n, prob = 0.4)
x <- rnorm(100)
data2 <- data.frame(n, success, x)
fit2 <- brm(success | trials(n) ~ x, data = data2,
            family = binomial("probit"))
summary(fit2)
```

--------------------------------

### Fit a brms model with OpenCL acceleration

Source: https://paulbuerkner.com/brms/reference/opencl.html

Example of fitting a brms model using OpenCL acceleration. Note that OpenCL may not provide speedups for all models or hardware configurations. Ensure `backend = "cmdstanr"` is specified.

```R
if (FALSE) {
# this model just serves as an illustration
# OpenCL may not actually speed things up here
fit <- brm(count ~ zAge + zBase * Trt + (1|patient),
           data = epilepsy, family = poisson(),
           chains = 2, cores = 2, opencl = opencl(c(0, 0)),
           backend = "cmdstanr")
summary(fit)
}
```

--------------------------------

### Example: Posterior Smooths Prediction

Source: https://paulbuerkner.com/brms/reference/posterior_smooths.brmsfit.html

Demonstrates how to fit a brms model with smooth terms and then compute posterior predictions for a specific smooth term using `posterior_smooths` with new data. Requires the `mgcv` package for data simulation.

```R
set.seed(0)
dat <- mgcv::gamSim(1, n = 200, scale = 2)
fit <- brm(y ~ s(x0) + s(x1) + s(x2) + s(x3), data = dat)
summary(fit)

newdata <- data.frame(x2 = seq(0, 1, 10))
str(posterior_smooths(fit, smooth = "s(x2)", newdata = newdata))
```
Tokens:
Raw