Negative binomial fit over millions of data

If your model is as simple as you write than MCMC sampling might be an overkill. MLE is probably going to give you the very same results

1 Like