I am trying to model a data with 6 repeats (number of rows) and 40 categories (number of columns) using multinomial distribution (see below for the data and code). I tried to model this data using two different shape arguments to the likelihood function which I thought would implicitly be broadcaste…

This may help: Distribution Dimensionality — PyMC 5.5.0 documentation Specially the multivariate examples

Confusion on the use of shape parameter in multinomial likelihood

iavicenna July 21, 2023, 12:36am 5

I can confirm (sorry for forgetting to reply) that they do behave as I expected when I used pymc instead of pymc3.

1 Like

Topic		Replies	Views
Fitting multiple measurements, shape issues in sample_ppc Questions	4	1448	November 8, 2018
Shape error when sampling more than one value Questions	13	1907	December 20, 2018
Help with dimension shape ( DirichletMultinomial ) v3	0	428	May 30, 2022
Using shape parameter vs using specifying it using a different model altogether Questions	5	2023	August 29, 2018
Dynamic number of Dirichlet distributions? Questions	3	622	February 25, 2018