mape_binomial: mape_binomial

In alexandercoulter/DHBayes: Bayesian Hierarchical Sampler for Discrete (Binomial, Negative-Binomial, Poisson) Data

Description

mape_binomial

Usage

mape_binomial(df, a_prior, b_prior)

Arguments

df	data.frame object, containing at least columns named 'x' containing non-negative integer values (number of successes), and 'n' containing non-negative integer values (number of trials).
a_prior	positive numeric, giving prior parameter 'a', assuming binomial parameter p ~ beta(a, b).
b_prior	positive numeric, giving prior parameter 'b', assuming binomial parameter p ~ beta(a, b).

Value

numeric, maximum a-posteriori estimate (MAPE) of binomial parameter p, assuming df$x ~ binom(p, df$n), and p ~ beta(a, b).

Examples

# Generate example data:
set.seed(31)
p = 0.3

# Number of experiments, i.e. rows in df:
numexps = 10

# Filling df with pseudo data; note the requisite columns 'n' and 'x':
n = 5 + rpois(numexps, 10)
x = rbinom(numexps, n, p)
df = data.frame('n' = n, 'x' = x)

# Generating maximum a posteriori estimate (MAPE) solution for p:
p_fit = mape_binomial(df = df, a_prior = 1, b_prior = 1)

# Compare fitted values to known values:
cbind(p, p_fit)

alexandercoulter/DHBayes documentation built on Dec. 19, 2021, 12:29 a.m.