# bayesplay: The Bayesian playground

The goal of bayesplay is to provide an interface for calculating Bayes factors for simple models. It does this in a way that makes the calculations more *transparent* and it is therefore useful as a teaching tools.

## Installation

bayesplay is now on CRAN. You can install it with:

`install.packages("bayesplay")`

Or if you want to live on the edge, you can install the development version from GitHub with:

```
# install.packages("devtools")
devtools::install_github("bayesplay/bayesplay")
```

## Basic usage

The `bayesplay`

package comes with three basic functions for computing Bayes factors.

The `likelihood()`

function for specifying likelihoods

The `prior()`

function for specifying priors

And the `integral()`

function

Currently the following distributions are supported for likelihoods and priors

### Priors

Normal distribution (`normal`

)

Uniform distribution (`uniform`

)

Scaled and shifted *t* distribution (`student_t`

)

Cauchy distributions (`cauchy`

)

Beta distribution (`beta`

)

### Likelihood

Normal distribution (`normal`

)

Scaled and shifted *t* distribution (`student_t`

)

Binomial distribution (`binomial`

)

Various noncentral *t* distributions, including:

Noncentral *t* distribution (`noncentral_t`

)

Noncentral *t* distribution scaled for a paired samples/one sample Cohen’s *d* (`noncentral_d`

)

Noncentral *t* distribution scaled for an independent samples Cohen’s *d* (`noncentral_d2`

)

## Worked examples

For worked examples of the basic usage see basic usage. Or for basic plot functionality see basic plotting

## Changelog

Breaking changes for < v0.9.0

`distribution`

parameter for specifying likelihoods and priors has been renamed `family`

`noncentral_d`

and `noncentral_d2`

are now parametrised in terms of sample size rather than df