Sample variables from a negative binomial distribution with fixed sum

This function samples from a negative binomial distribution. Rather than specifying the total number of samples, as with rnbinom, this function takes as input an integer N which the resulting samples must sum to. This is achieved by adjusting the final sample. As such, the resulting samples are not strictly speaking from a negative binomial distribution.

Usage

sample_negbinom(N, prop_max = 0.1, mu, size)

Arguments

N

An integer giving the sum of generated random variables

prop_max

The proportion of N which is the maximum value of a single sample

mu

See the mu argument of rnbinom()

size

See the size argument of rnbinom()

See also

Other sampling: sample_log_normal(), sample_offset_truncated_power_distribution()