mBayesBinom: Interactive applet to explore Bayesian estimates

In mosaicManip: Project MOSAIC (mosaic-web.org) manipulate examples.

Description

An applet which lets the user manipulate the prior distribution and sample size to explore its effect on the posterior distribution and Bayesian estimates.

Usage

mBayesBinom(dat, ...)

Arguments

dat	Binomial data consisting of zeros and ones to be used as the likelihood.
...	Extra arguments will not be evaluated.

Details

The user specifies the prior distribution to use using the corresponding picker. The beta distribution is the default as it is smooth and more applicable to real life data, where our prior belief would drop exponentially as distance increases from the mean, but it would not equal zero. When using the beta distribution, the parameter 1 slider controls alpha, the parameter 2 slider controls beta. When using the uniform distribution as the prior, parameter 1 controls the minimum and parameter 2 controls the maximum. Sample size allows the user to sample any number of points from one point to twice the size of the data set (the data set repeated).

Known bugs: Logarithmic y-axis checkbox does not work, unfortunately.

Value

A function that allows the user to explore bayesian inference interactively.

Author(s)

Andrew Rich ([e-mail]) and Daniel Kaplan ([e-mail])

Examples

	if(require(manipulate)) {
		data(CPS)
		mBayesBinom(CPS$sex)
	}

mosaicManip documentation built on May 2, 2019, 5:48 p.m.