bayesplain.core.gamma_poisson.log_marginal_likelihood

bayesplain.core.gamma_poisson.log_marginal_likelihood(events, exposure, a=1.0, b=1.0)[source]

Log marginal likelihood of a count under a proper Gamma prior.

Parameters:
events : int

Events counted.

exposure : float

Exposure.

a : float

Prior shape and rate parameter. b must be positive: an improper prior has no marginal likelihood.

b : float

Prior shape and rate parameter. b must be positive: an improper prior has no marginal likelihood.

Returns:

Log of the negative-binomial probability of the observed count.

Return type:

float