bayestestR

https://easystats.github.io/bayestestR/
Admin
bayestestR provides a comprehensive set of functions to analyze and describe posterior distributions...
Tokens:126,197
Snippets:1,145
Trust Score:9.7
Update:5 months ago
Show doc for...
Tokens:
Raw
### Install easystats and bayestestR

Source: https://easystats.github.io/bayestestR/articles/bayestestR.html

Install the easystats suite, which includes bayestestR, from GitHub. This requires the 'remotes' package.

```r
install.packages("remotes")
remotes::install_github("easystats/easystats")
```

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

### Install and load rstanarm

Source: https://easystats.github.io/bayestestR/articles/bayestestR.html

Install the 'rstanarm' package for fitting Bayesian models and load it along with bayestestR for model description.

```r
install.packages("rstanarm")
library(rstanarm)
```

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

### Example: Bayesian Model with Equal Priors

Source: https://easystats.github.io/bayestestR/reference/contr.equalprior.html

Demonstrates setting equal priors on dummy variables in a Bayesian model using `brms`. This example contrasts the default `contr.treatment` with the approach using custom contrast functions to achieve marginally equal priors on means and their differences.

```R
library(brms)

data <- data.frame(
  group = factor(rep(LETTERS[1:4], each = 3)),
  y = rnorm(12)
)

contrasts(data$group) # R's default contr.treatment
#>   B C D
#> A 0 0 0
#> B 1 0 0
#> C 0 1 0
#> D 0 0 1

model_prior <- brm(
  y ~ group, data = data,
  sample_prior = "only",
  # Set the same priors on the 3 dummy variable
  # (Using an arbitrary scale)
  prior = set_prior("normal(0, 10)", coef = c("groupB", "groupC", "groupD"))
)

est <- emmeans::emmeans(model_prior, pairwise ~ group)

point_estimate(est, centr = "mean", disp = TRUE)
#> Point Estimate
#>
#> Parameter |  Mean |    SD
#> -------------------------
#> A         | -0.01 |  6.35
#> B         | -0.10 |  9.59
#> C         |  0.11 |  9.55
#> D         | -0.16 |  9.52
#> A - B     |  0.10 |  9.94
#> A - C     | -0.12 |  9.96
#> A - D     |  0.15 |  9.87
#> B - C     | -0.22 | 14.38
#> B - D     |  0.05 | 14.14
#> C - D     |  0.27 | 14.00
```
```

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

### Calculate MAP Estimate from a brms Model

Source: https://easystats.github.io/bayestestR/reference/map_estimate.html

This example shows how to obtain the MAP estimate from a model fitted using the `brms` package. Ensure the `brms` package is installed and loaded.

```R
model <- brms::brm(mpg ~ wt + cyl, data = mtcars)
map_estimate(model)
```

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

### Describe Posterior with All Options

Source: https://easystats.github.io/bayestestR/reference/describe_posterior.html

This example shows how to use `describe_posterior` with arguments to include all centrality measures, dispersion statistics, and test types for a thorough analysis.

```r
describe_posterior(bf, centrality = "all", dispersion = TRUE, test = "all")
```

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

### Load Libraries and Data

Source: https://easystats.github.io/bayestestR/articles/web_only/indicesEstimationComparison.html

Loads necessary R libraries and reads simulation data for analysis. Ensure these libraries are installed.

```r
library(ggplot2)
library(datawizard)
library(see)
library(parameters)

df <- read.csv("https://raw.github.com/easystats/circus/main/data/bayesSim_study1.csv")
```

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

### ROPE Proportion of Samples Results

Source: https://easystats.github.io/bayestestR/articles/region_of_practical_equivalence.html

Example output showing the proportion of posterior samples falling within the default ROPE range for each parameter.

```text
> # Proportion of samples inside the ROPE [-0.08, 0.08]:
> 
> Parameter   | Inside ROPE
> -------------------------
> (Intercept) |      0.00 %
> Sepal.Width |     16.28 %
```

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

### Probability of Direction Results

Source: https://easystats.github.io/bayestestR/articles/region_of_practical_equivalence.html

Example output showing the Probability of Direction (pd) for each parameter in the model.

```text
> Probability of Direction
> 
> Parameter   |     pd
> --------------------
> (Intercept) |   100%
> Sepal.Width | 91.65%
```

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

### Load R Packages for Bayesian Modeling

Source: https://easystats.github.io/bayestestR/articles/example1.html

Loads the 'rstanarm', 'bayestestR', and 'insight' packages. Ensure these are installed before running.

```R
library(rstanarm)
library(bayestestR)
library(insight)
```

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

### Example: Reshaping Posterior Draws

Source: https://easystats.github.io/bayestestR/reference/reshape_iterations.html

This example demonstrates how to reshape posterior draws obtained from a `stan_glm` model into a long format using `reshape_iterations`. It requires the `rstanarm` and `insight` packages. The `head()` function shows the first few rows of the reshaped data.

```R
# \donttest{
if (require("rstanarm")) {
  model <- stan_glm(mpg ~ am, data = mtcars, refresh = 0)
  draws <- insight::get_predicted(model)
  long_format <- reshape_iterations(draws)
  head(long_format)
}
# }
```

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

### Simulate Prior from rstanarm Model

Source: https://easystats.github.io/bayestestR/reference/simulate_prior.html

This snippet shows how to use `simulate_prior` with a model fitted using `stan_glm` from the `rstanarm` package. Ensure `rstanarm` is installed and loaded.

```R
library(bayestestR)
if (require("rstanarm")) {
  model <- suppressWarnings(
    stan_glm(mpg ~ wt + am, data = mtcars, chains = 1, refresh = 0)
  )
  simulate_prior(model)
}
```

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

### Visualize MAP Estimate on Posterior Density Plot

Source: https://easystats.github.io/bayestestR/reference/map_estimate.html

This example shows how to plot the density of a posterior distribution and overlay the calculated MAP estimate as a vertical red line.

```R
plot(density(posterior))
abline(v = as.numeric(map_estimate(posterior)), col = "red")
```

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

### Load Libraries for bayestestR and ggplot2

Source: https://easystats.github.io/bayestestR/articles/credible_interval.html

Load the necessary libraries for statistical analysis and visualization. Ensure these packages are installed before running.

```r
library(bayestestR)
library(ggplot2)
```

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

### Simulate Multiple Correlated Variables

Source: https://easystats.github.io/bayestestR/reference/simulate_correlation.html

Creates a dataset with multiple variables exhibiting a specified correlation matrix. This example shows how to define a correlation matrix and use it with `simulate_correlation` to generate data for multivariate analysis.

```R
cor_matrix <- matrix(
  c(
    1.0, 0.2, 0.4,
    0.2, 1.0, 0.3,
    0.4, 0.3, 1.0
  ),
  nrow = 3
)

data <- simulate_correlation(r = cor_matrix, names = c("y", "x1", "x2"))
cor(data)
summary(lm(y ~ x1, data = data))
```

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

### Calculate Specific Point Estimates for a Data Frame

Source: https://easystats.github.io/bayestestR/reference/point_estimate.html

This example shows how to compute specific point estimates (median and MAP) for each column in a data frame by specifying them in the `centrality` argument.

```R
point_estimate(df, centrality = c("median", "MAP"))
```

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

### Calculate Support Interval with BF = 3

Source: https://easystats.github.io/bayestestR/reference/si.html

This example shows how to use the `si` function to calculate the Support Interval with a specified Bayes Factor threshold of 3. This allows for a more stringent criterion for evidence.

```r
si(stan_model, BF = 3, verbose = FALSE)
```

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

### Calculate All Point Estimates with Dispersion for a Distribution

Source: https://easystats.github.io/bayestestR/reference/point_estimate.html

This example shows how to compute all available point estimates, including dispersion measures like MAD and SD, for a numeric vector. The `centrality` argument is set to 'all' and `dispersion` to TRUE.

```R
point_estimate(rnorm(1000), centrality = "all", dispersion = TRUE)
```

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

### ROPE Proportion of Samples Results with Scaled Predictor

Source: https://easystats.github.io/bayestestR/articles/region_of_practical_equivalence.html

Example output showing the proportion of posterior samples within the ROPE for the scaled predictor. This demonstrates how scaling drastically affects ROPE interpretation.

```text
> # Proportion of samples inside the ROPE [-0.08, 0.08]:
> 
> Parameter          | Inside ROPE
> --------------------------------
> (Intercept)        |      0.00 %
> Sepal.Width_scaled |      0.10 %
```

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

### Describe Prior for rstanarm Model

Source: https://easystats.github.io/bayestestR/reference/describe_prior.html

Fit a generalized linear model using rstanarm and then visualize its prior distributions. Requires the rstanarm package to be installed.

```R
if (require("rstanarm")) {
  model <- rstanarm::stan_glm(mpg ~ wt + cyl, data = mtcars)
  describe_prior(model)
}
```

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

### Load and Display 'disgust' Dataset in R

Source: https://easystats.github.io/bayestestR/reference/disgust.html

Loads the 'disgust' dataset and displays the first 5 rows. This dataset is used for tests and examples.

```r
data("disgust")
head(disgust, n = 5)
```

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

### Calculate Specific Point Estimates for a Distribution

Source: https://easystats.github.io/bayestestR/reference/point_estimate.html

To calculate only specific point estimates, provide a character vector to the `centrality` argument. This example calculates the median and MAP.

```R
point_estimate(rnorm(1000), centrality = c("median", "MAP"))
```

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

### Calculate Default Point Estimates for a Distribution

Source: https://easystats.github.io/bayestestR/reference/point_estimate.html

Use this to quickly get the median, mean, and MAP for a numeric vector. Ensure the `bayestestR` library is loaded.

```R
library(bayestestR)

point_estimate(rnorm(1000))
```

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

### Describe Prior for a brms Model

Source: https://easystats.github.io/bayestestR/reference/describe_prior.html

This snippet demonstrates how to use the describe_prior function with a model fitted using the brms package. Ensure the brms package is installed and loaded. The output shows a table summarizing the priors for each parameter.

```R
if (require("brms")) {
  model <- brms::brm(mpg ~ wt + cyl, data = mtcars)
  describe_prior(model)
}
```

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

### Convert brms Model Posteriors to Priors

Source: https://easystats.github.io/bayestestR/reference/model_to_priors.html

Demonstrates converting posterior distributions from a brms model to priors. The resulting priors are then validated and used to refit the model. Ensure the 'brms' package is installed and loaded.

```R
# \donttest{
# brms models
# -----------------------------------------------
if (require("brms")) {
  formula <- brms::brmsformula(mpg ~ wt + cyl, center = FALSE)

  model <- brms::brm(formula, data = mtcars, refresh = 0)
  priors <- model_to_priors(model)
  priors <- brms::validate_prior(priors, formula, data = mtcars)
  priors

  model2 <- brms::brm(formula, data = mtcars, prior = priors, refresh = 0)
}
#> Compiling Stan program...
#> Start sampling
#> Compiling Stan program...
#> Start sampling
# }
```

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

### Describe Priors for Bayesian Models

Source: https://easystats.github.io/bayestestR/reference/describe_prior.html

Use this function to get a summary of the priors in your Bayesian model. Meta-parameters are filtered by default. You can select specific parameters using the 'parameters' argument.

```R
library(bayestestR)

# rstanarm models

```

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

### Compare lmerMod Models with bayesfactor_models

Source: https://easystats.github.io/bayestestR/reference/bayesfactor_models.html

Compares mixed-effects models (`lmerMod`) using Bayes Factors. This example demonstrates comparing models with different random effects structures. The `estimator = "REML"` argument specifies the estimation method.

```R
lmer1 <- lme4::lmer(Sepal.Length ~ Petal.Length + (1 | Species), data = iris)
lmer2 <- lme4::lmer(Sepal.Length ~ Petal.Length + (Petal.Length | Species), data = iris)
lmer3 <- lme4::lmer(
  Sepal.Length ~ Petal.Length + (Petal.Length | Species) + (1 | Petal.Width),
  data = iris
)
bayesfactor_models(lmer1, lmer2, lmer3,
  denominator = 1,
  estimator = "REML"
)
```

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

### Define and Run SEM for Mediation in lavaan

Source: https://easystats.github.io/bayestestR/articles/mediation.html

Use this code to define a mediation model with direct, indirect, and total effects in lavaan. Ensure the 'lavaan' package is installed and loaded. The 'summary()' function displays the model results.

```r
library(lavaan)
data(jobs)
set.seed(1234)

model <- " # direct effects
            depress2 ~ c1*treat + c2*econ_hard + c3*sex + c4*age + b*job_seek

          # mediation
            job_seek ~ a1*treat + a2*econ_hard + a3*sex + a4*age

          # indirect effects (a*b)
            indirect_treat := a1*b
            indirect_econ_hard := a2*b
            indirect_sex := a3*b
            indirect_age := a4*b

          # total effects
            total_treat := c1 + (a1*b)
            total_econ_hard := c2 + (a2*b)
            total_sex := c3 + (a3*b)
            total_age := c4 + (a4*b)
        "
m4 <- sem(model, data = jobs)
summary(m4)
```

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

### distribution_beta()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample from a Beta distribution.

```APIDOC
## distribution_beta(n, shape1, shape2, ncp = 0, random = FALSE, ...)

### Description
Generates a sample from a Beta distribution.

### Arguments
* `n` (number) - Number of observations.
* `shape1` (numeric) - Non-negative parameter of the Beta distribution.
* `shape2` (numeric) - Non-negative parameter of the Beta distribution.
* `ncp` (numeric) - Non-centrality parameter. Defaults to 0.
* `random` (boolean) - If `FALSE`, generates a near-perfect distribution. If `TRUE`, generates a random sample. Defaults to `FALSE`.
* `...` - Additional arguments.
```

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

### Calculate PD for Data Frame

Source: https://easystats.github.io/bayestestR/reference/p_direction.html

Use p_direction with a data frame to get the probability of direction for each column. The default method is used.

```r
df <- data.frame(replicate(4, rnorm(100)))
p_direction(df)
```

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

### Visualize Bayes Factor with Pizza Chart

Source: https://easystats.github.io/bayestestR/articles/example2.html

Generates a 'pizza chart' visualization for Bayes Factors using the 'see' package. This provides an intuitive interpretation of the strength of evidence.

```r
library(see)

plot(bayesfactor_models(result)) +
  scale_fill_pizza()
```

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

### Compare Density Estimation Methods with p_map

Source: https://easystats.github.io/bayestestR/reference/p_map.html

This code snippet demonstrates how to compare the robustness of different density estimation methods ('kernel', 'KernSmooth', 'logspline') using the p_map function. It simulates data, applies each method, and then analyzes the differences.

```R
set.seed(333)
data <- data.frame()
for (iteration in 1:250) {
  x <- rnorm(1000, 1, 1)
  result <- data.frame(
    Kernel = as.numeric(p_map(x, method = "kernel")),
    KernSmooth = as.numeric(p_map(x, method = "KernSmooth")),
    logspline = as.numeric(p_map(x, method = "logspline"))
  )
  data <- rbind(data, result)
}
data$KernSmooth <- data$Kernel - data$KernSmooth
data$logspline <- data$Kernel - data$logspline

summary(data$KernSmooth)
summary(data$logspline)
boxplot(data[c("KernSmooth", "logspline")])
```

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

### Calculate Point Estimates from emmeans

Source: https://easystats.github.io/bayestestR/reference/point_estimate.html

Use `point_estimate` with `emtrends` from the `emmeans` package to get median and MAP estimates. Ensure the `emmeans` package is loaded.

```r
point_estimate(
  emmeans::emtrends(model, ~1, "wt", data = mtcars),
  centrality = c("median", "MAP")
)
```

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

### Generate Near-Perfect Distributions with `distribution()`

Source: https://easystats.github.io/bayestestR/index.html

Generates a sample of size n with near-perfect distributions. Use this function to create samples that closely follow a specific distribution.

```R
distribution(n = 10)
##  [1] -1.55 -1.00 -0.66 -0.38 -0.12  0.12  0.38  0.66  1.00  1.55
```

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

### Visualize Warmup Ratio Sensitivity with ggplot2

Source: https://easystats.github.io/bayestestR/articles/web_only/indicesEstimationComparison.html

Prepares and visualizes data to show how different estimation indices (beta, MAP, Mean, Median) are affected by the warmup ratio. Requires the ggplot2 library for plotting.

```r
dat <- df
dat$warmup <- dat$warmup / dat$iterations
dat <- data_select(dat, select = c("warmup", "true_effect", "outcome_type", "beta", "Median", "Mean", "MAP"))
dat <- reshape_longer(
  dat,
  select = -c("warmup", "true_effect", "outcome_type"),
  names_to = "estimate",
  values_to = "value"
)
dat$temp <- as.factor(cut(dat$warmup, 3, labels = FALSE))

tmp <- lapply(split(dat, dat$temp), function(x) {
  x$warmup_group <- rep(round(mean(x$warmup), 1), times = nrow(x))
  return(x)
})

dat <- do.call(rbind, tmp)
dat <- data_filter(dat, value < 6)

ggplot(dat, aes(x = warmup_group, y = value, fill = estimate, group = interaction(estimate, warmup_group))) +
  geom_boxplot(outlier.shape = NA) +
  theme_classic() +
  scale_fill_manual(
    values = c("beta" = "#607D8B", "MAP" = "#795548", "Mean" = "#FF9800", "Median" = "#FFEB3B"),
    name = "Index"
  ) +
  ylab("Point-estimate of the true value 0\n") +
  xlab("\nNumber of Iterations") +
  facet_wrap(~ outcome_type * true_effect, scales = "free")
```

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

### Fit Unrestricted GLM Model with Priors

Source: https://easystats.github.io/bayestestR/articles/bayes_factors.html

Fits a generalized linear model using Stan with specified priors. Ensure the `rstanarm` package is installed and loaded.

```r
iris_model <- stan_glm(Sepal.Length ~ Species + Petal.Length,
  data = iris,
  prior = normal(0, c(2, 1.2, 1.2), autoscale = FALSE),
  chains = 10, iter = 5000, warmup = 1000
)
```

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

### Calculate sexit thresholds for brms models

Source: https://easystats.github.io/bayestestR/reference/sexit_thresholds.html

Apply `sexit_thresholds` to models fitted using the `brms` package. Ensure the `brms` package is installed and loaded before use.

```R
if (require("brms")) {
  model <- brm(mpg ~ wt + cyl, data = mtcars)
  sexit_thresholds(model)
}
```

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

### distribution()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample of size `n` from a specified distribution. By default, it generates a normal distribution. Additional arguments can be passed to customize the distribution.

```APIDOC
## distribution(type = "normal", ...)

### Description
Generates a sample of size `n` from a specified distribution. By default, it generates a normal distribution. Additional arguments can be passed to customize the distribution.

### Arguments
* `type` (string) - The type of distribution to generate. Can be any of the names from base R's Distributions, like `"cauchy"`, `"pois"` or `"beta"`.
* `...` - Arguments passed to or from other methods.

### Examples
```R
library(bayestestR)
x <- distribution(n = 10)
plot(density(x))

x <- distribution(type = "gamma", n = 100, shape = 2)
plot(density(x))
```
```

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

### Describe Prior Distribution of a Bayes Factor Object

Source: https://easystats.github.io/bayestestR/reference/describe_prior.html

Use `describe_prior` to inspect the prior distribution of a Bayes factor object. This requires the BayesFactor package to be installed and loaded.

```R
if (require("BayesFactor")) {
  bf <- ttestBF(x = rnorm(100, 1, 1))
  describe_prior(bf)
}
#>    Parameter Prior_Distribution Prior_Location Prior_Scale
#> 1 Difference             cauchy              0   0.7071068
```

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

### distribution_tweedie()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample from a Tweedie distribution.

```APIDOC
## distribution_tweedie(n, xi = NULL, mu, phi, power = NULL, random = FALSE, ...)

### Description
Generates a sample from a Tweedie distribution.

### Arguments
* `n` (number) - Number of observations.
* `xi` (numeric) - For tweedie distributions, the value of `xi` such that the variance is `var(Y) = phi * mu^xi`.
* `mu` (numeric) - Vector of mean μ̂.
* `phi` (numeric) - Variance scale parameter.
* `power` (numeric) - Alias for `xi`.
* `random` (boolean) - If `FALSE`, generates a near-perfect distribution. If `TRUE`, generates a random sample. Defaults to `FALSE`.
* `...` - Additional arguments.
```

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

### distribution_cauchy()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample from a Cauchy distribution.

```APIDOC
## distribution_cauchy(n, location = 0, scale = 1, random = FALSE, ...)

### Description
Generates a sample from a Cauchy distribution.

### Arguments
* `n` (number) - Number of observations.
* `location` (numeric) - Location parameter. Defaults to 0.
* `scale` (numeric) - Scale parameter. Defaults to 1.
* `random` (boolean) - If `FALSE`, generates a near-perfect distribution. If `TRUE`, generates a random sample. Defaults to `FALSE`.
* `...` - Additional arguments.
```

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

### Calculate sexit thresholds for rstanarm models

Source: https://easystats.github.io/bayestestR/reference/sexit_thresholds.html

Use `sexit_thresholds` with models fitted by `stan_glm` from the `rstanarm` package. This example demonstrates its use with both linear and binomial regression models.

```R
if (require("rstanarm")) {
  model <- suppressWarnings(stan_glm(
    mpg ~ wt + gear,
    data = mtcars,
    chains = 2,
    iter = 200,
    refresh = 0
  ))
  sexit_thresholds(model)

  model <- suppressWarnings(
    stan_glm(vs ~ mpg, data = mtcars, family = "binomial", refresh = 0)
  )
  sexit_thresholds(model)
}
```

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

### Probability of Direction Results with Scaled Predictor

Source: https://easystats.github.io/bayestestR/articles/region_of_practical_equivalence.html

Example output for Probability of Direction (pd) after scaling the predictor. The pd value remains similar, but the ROPE interpretation changes.

```text
> Probability of Direction
> 
> Parameter          |     pd
> ---------------------------
> (Intercept)        |   100%
> Sepal.Width_scaled | 91.65%
```

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

### Basic Density Estimation (Kernel Method)

Source: https://easystats.github.io/bayestestR/reference/estimate_density.html

Estimates the density of a numeric vector using the default kernel method and visualizes the estimate with its confidence interval. Requires the bayestestR library.

```R
library(bayestestR)

set.seed(1)
x <- rnorm(250, mean = 1)

# Basic usage
density_kernel <- estimate_density(x) # default method is "kernel"

hist(x, prob = TRUE)
lines(density_kernel$x, density_kernel$y, col = "black", lwd = 2)
lines(density_kernel$x, density_kernel$CI_low, col = "gray", lty = 2)
lines(density_kernel$x, density_kernel$CI_high, col = "gray", lty = 2)
legend("topright",
  legend = c("Estimate", "95% CI"),
  col = c("black", "gray"), lwd = 2, lty = c(1, 2)
)
```

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

### distribution_student() / distribution_t() / distribution_student_t()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample from a Student's t-distribution.

```APIDOC
## distribution_student(n, df, ncp, random = FALSE, ...)
## distribution_t(n, df, ncp, random = FALSE, ...)
## distribution_student_t(n, df, ncp, random = FALSE, ...)

### Description
Generates a sample from a Student's t-distribution.

### Arguments
* `n` (number) - Number of observations.
* `df` (numeric) - Degrees of freedom.
* `ncp` (numeric) - Non-centrality parameter.
* `random` (boolean) - If `FALSE`, generates a near-perfect distribution. If `TRUE`, generates a random sample. Defaults to `FALSE`.
* `...` - Additional arguments.
```

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

### Summarize Posterior Distribution from rstanarm Model

Source: https://easystats.github.io/bayestestR/reference/describe_posterior.html

Use this to get a basic summary of the posterior distribution for parameters in an rstanarm model. Ensure the model is fitted before calling this function.

```r
model <- suppressWarnings(
  rstanarm::stan_glm(
    mpg ~ wt + gear,
    data = mtcars, chains = 2, iter = 200,
    refresh = 0
  )
)
describe_posterior(model)
```

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

### Get Diagnostic Draws from MCMC Models

Source: https://easystats.github.io/bayestestR/reference/diagnostic_draws.html

Use diagnostic_draws to extract MCMC diagnostic information. Supports stanreg, stanfit, brmsfit, blavaan, or raw MCMC data.

```r
diagnostic_draws(posterior, ...)
```

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

### Get Number of Posterior Draws

Source: https://easystats.github.io/bayestestR/articles/example1.html

Use `nrow(posteriors)` to determine the number of posterior draws. This count is crucial for understanding the precision of your Bayesian model's estimation.

```R
nrow(posteriors) # Size (number of rows)
```

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

### Fit a Bayesian Linear Model with rstanarm

Source: https://easystats.github.io/bayestestR/reference/sensitivity_to_prior.html

Fit a Bayesian linear regression model using `stan_glm` from the `rstanarm` package. This is a common starting point for Bayesian analysis.

```r
model <- rstanarm::stan_glm(mpg ~ wt, data = mtcars)
```

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

### distribution_binomial() / distribution_binom()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample from a Binomial distribution.

```APIDOC
## distribution_binomial(n, size = 1, prob = 0.5, random = FALSE, ...)
## distribution_binom(n, size = 1, prob = 0.5, random = FALSE, ...)

### Description
Generates a sample from a Binomial distribution.

### Arguments
* `n` (number) - Number of observations.
* `size` (integer) - Number of trials (zero or more). Defaults to 1.
* `prob` (numeric) - Probability of success on each trial. Defaults to 0.5.
* `random` (boolean) - If `FALSE`, generates a near-perfect distribution. If `TRUE`, generates a random sample. Defaults to `FALSE`.
* `...` - Additional arguments.
```

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

### Load bayestestR Package

Source: https://easystats.github.io/bayestestR/index.html

Load the bayestestR package for use. Alternatively, loading 'easystats' makes all easystats ecosystem features available.

```r
library("bayestestR")
```

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

### Calculate ETI for emtrends with rstanarm model

Source: https://easystats.github.io/bayestestR/reference/eti.html

Apply `eti` to the output of `emmeans::emtrends` to calculate the Equal-Tailed Interval for estimated marginal trends, specifically for the 'wt' variable in this example.

```r
eti(emmeans::emtrends(model, ~1, "wt", data = mtcars))
```

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

### distribution_uniform()

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample from a Uniform distribution.

```APIDOC
## distribution_uniform(n, min = 0, max = 1, random = FALSE, ...)

### Description
Generates a sample from a Uniform distribution.

### Arguments
* `n` (number) - Number of observations.
* `min` (numeric) - Lower limit of the distribution. Must be finite. Defaults to 0.
* `max` (numeric) - Upper limit of the distribution. Must be finite. Defaults to 1.
* `random` (boolean) - If `FALSE`, generates a near-perfect distribution. If `TRUE`, generates a random sample. Defaults to `FALSE`.
* `...` - Additional arguments.
```

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

### Calculate ROPE for brms model (mpg ~ wt + cyl)

Source: https://easystats.github.io/bayestestR/reference/rope_range.html

Fit a Bayesian regression model using brms and then calculate the ROPE. This example includes the full sampling output, which can be lengthy.

```r
model <- brms::brm(mpg ~ wt + cyl, data = mtcars)
rope_range(model)
```

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

### Load Simulation Data

Source: https://easystats.github.io/bayestestR/articles/web_only/indicesEstimationComparison.html

Loads the dataset for Experiment 2, which contains simulation results for Bayesian and frequentist models.

```R
df <- read.csv("https://raw.github.com/easystats/circus/main/data/bayesSim_study2.csv")
```

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

### Get Pairwise Bayes Factor Matrix

Source: https://easystats.github.io/bayestestR/articles/bayes_factors.html

Converts a model comparison object into a matrix format, showing pairwise Bayes factors between all compared models. Useful for a comprehensive overview of all comparisons.

```R
as.matrix(comparison)
```

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

### Load Data and Fit Model with BayesFactor Package

Source: https://easystats.github.io/bayestestR/articles/bayes_factors.html

Loads the ToothGrowth dataset and fits a Bayesian linear model using the `BayesFactor` package for subsequent Bayes factor calculations.

```R
library(BayesFactor)
data(ToothGrowth)
ToothGrowth$dose <- as.factor(ToothGrowth$dose)

BF_ToothGrowth <- anovaBF(len ~ dose * supp, ToothGrowth, progress = FALSE)
```

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

### Summarize Posterior Distribution

Source: https://easystats.github.io/bayestestR/reference/describe_posterior.html

Use this to get a basic summary of the posterior distribution, including median, 95% credible interval, probability of direction (pd), and region of practical equivalence (ROPE).

```r
bf <- BayesFactor::ttestBF(x = rnorm(100, 1, 1))
describe_posterior(bf)
```

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

### Running check_prior with Gelman Method

Source: https://easystats.github.io/bayestestR/reference/check_prior.html

This snippet shows how to use `check_prior` with the 'gelman' method to assess prior quality for a fitted `stan_glm` model. It requires the `rstanarm` package and a model object.

```r
model <- rstanarm::stan_glm(mpg ~ wt,
  data = mtcars[1:3, ],
  prior = normal(-3.3, 1, FALSE),
  prior_intercept = normal(0, 1000, FALSE),
  refresh = 0
)
check_prior(model, method = "gelman")
```

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

### Generate a Gamma Distribution Sample

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample of 100 observations from a gamma distribution with a shape parameter of 2 and plots its density. Requires the bayestestR library.

```R
x <- distribution(type = "gamma", n = 100, shape = 2)
plot(density(x))
```

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

### Running check_prior with Lakeland Method

Source: https://easystats.github.io/bayestestR/reference/check_prior.html

This snippet demonstrates using `check_prior` with the 'lakeland' method. This method provides an alternative assessment of prior quality for a `stan_glm` model.

```r
check_prior(model, method = "lakeland")
```

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

### Calculate MAP-based p-values for emmeans object

Source: https://easystats.github.io/bayestestR/reference/p_map.html

Apply p_map to an emtrends object from the emmeans package to get MAP-based p-values for estimated marginal trends. This is useful for interpreting contrasts or specific comparisons.

```R
p_map(suppressWarnings(
  emmeans::emtrends(model, ~1, "wt", data = mtcars)
))
```

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

### Density Estimation with Extended Range

Source: https://easystats.github.io/bayestestR/reference/estimate_density.html

Demonstrates the effect of the 'extend' argument on density estimation. 'extend = TRUE' extends the density range, while 'extend = FALSE' uses the default range.

```R
# Extension
density_extended <- estimate_density(x, extend = TRUE)
density_default <- estimate_density(x, extend = FALSE)

hist(x, prob = TRUE)
lines(density_extended$x, density_extended$y, col = "red", lwd = 3)
lines(density_default$x, density_default$y, col = "black", lwd = 3)
```

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

### Fit Bayesian Logistic Model

Source: https://easystats.github.io/bayestestR/articles/example2.html

Fits a Bayesian logistic regression model using the rstanarm package. This model predicts the probability of a species based on Sepal.Width. Ensure 'data' is loaded and 'rstanarm' is installed.

```r
library(rstanarm)

model <- stan_glm(
  Species ~ Sepal.Width,
  data = data,
  family = "binomial",
  chains = 10,
  iter = 5000,
  warmup = 1000,
  refresh = 0
)
```

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

### Generate a Normal Distribution Sample

Source: https://easystats.github.io/bayestestR/reference/distribution.html

Generates a sample of 10 observations from a normal distribution and plots its density. Requires the bayestestR library.

```R
library(bayestestR)
x <- distribution(n = 10)
plot(density(x))
```

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

### Getting ROPE Range from Model

Source: https://easystats.github.io/bayestestR/articles/example1.html

Retrieves a default ROPE range, typically based on a fraction of the standard deviation of the model's response variable. This offers a standardized approach to defining negligible effects.

```r
rope_value <- rope_range(model)
rope_value
```

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

### Estimate Contrasts with Bayes Factor using `modelbased`

Source: https://easystats.github.io/bayestestR/articles/bayes_factors.html

Employ `estimate_contrasts` from the `modelbased` package to test contrasts using Bayes factors. Specify the prior model using `bf_prior`.

```R
library(modelbased)

estimate_contrasts(model, test = "bf", bf_prior = model)
```
Tokens:
Raw