How does pymc3 handle n=0 in Binomial discrete likelihood function

ricardoV94 · March 14, 2022, 3:03pm

But it doesn’t seem necessary I think. We could just modify these lines to include 0 <= n instead of 0 < n:

pymc-devs/pymc/blob/a6e6748e176a4bc14fff68ad69b7c7b418c94daf/pymc/distributions/discrete.py#L149

      
        
                -------
                TensorVariable
                """
            
            
    res = at.switch(
                    at.or_(at.lt(value, 0), at.gt(value, n)),
                    -np.inf,
                    binomln(n, value) + logpow(p, value) + logpow(1 - p, n - value),
                )
            
            
    return check_parameters(res, 0 < n, 0 <= p, p <= 1, msg="n > 0, 0 <= p <= 1")
            
            
def logcdf(value, n, p):
                """
                Compute the log of the cumulative distribution function for Binomial distribution
                at the specified value.
            
            
    Parameters
                ----------
                value: numeric or np.ndarray or aesara.tensor
                    Value(s) for which log CDF is calculated. If the log CDF for multiple

github.com

pymc-devs/pymc/blob/a6e6748e176a4bc14fff68ad69b7c7b418c94daf/pymc/distributions/discrete.py#L180

      
        
                        -np.inf,
                        at.switch(
                            at.lt(value, n),
                            at.log(at.betainc(n - value, value + 1, 1 - p)),
                            0,
                        ),
                    )
            
            
        return check_parameters(
                        res,
                        0 < n,
                        0 <= p,
                        p <= 1,
                        msg="n > 0, 0 <= p <= 1",
                    )
            
            

            
class BetaBinomial(Discrete):
                R"""
                Beta-binomial log-likelihood.

Topic		Replies	Views
Why does my model show logp of zero? Questions	3	592	June 12, 2020
Negative Binomial Distribution with logp = -inf Questions	14	1567	July 11, 2018
Sampler issues on Beta prior Binomial likelihood v5 bug	5	409	December 8, 2022
Beta Binomial distribution Questions	1	700	August 26, 2018
Get nan or inf from model.logp (model.test point) is an attestation of incorrectly configured model? Questions	22	8333	June 24, 2017

How does pymc3 handle n=0 in Binomial discrete likelihood function

Related topics