There are a variety of parameterizations of the negative binomial. The parameterizations I use throughout these slides are consistent with Bayesian Data Analysis (3rd edition) by Gelman, et. al. (see page 578). This parameterization uses the shape and rate (or inverse scale) and thus the expectation for the NB is shape divided by the rate, i.e. (alpha+n*ybar)/(beta+n). The gamma distribution also uses the shape and rate and has the same expectation. Of course, the two distributions have different supports.
preparing for lecture tomorrow
Do you have an email address that i could contact you? And would i be able to get the lecture notes??
Is the parameterization of the negative binomial in slide 10 correct? It looks perhaps like a mistaken copy of the parameterization of the gamma...
There are a variety of parameterizations of the negative binomial. The parameterizations I use throughout these slides are consistent with Bayesian Data Analysis (3rd edition) by Gelman, et. al. (see page 578). This parameterization uses the shape and rate (or inverse scale) and thus the expectation for the NB is shape divided by the rate, i.e. (alpha+n*ybar)/(beta+n). The gamma distribution also uses the shape and rate and has the same expectation. Of course, the two distributions have different supports.
Interesujace podejscie