So, basically when i use posterior predictive distribution for logistic regression, when number of points in test set is greater then 1, posterior predictive shape is as acting as we would expect. But, when it is exactly 1, posterior predictive shape becomes awkward? Would you check code example bel…

Weird posterior predictive shape when only one value is in test set?

OriolAbril February 22, 2021, 4:44pm 2

Yes, this is a know issue (ticket 3640) that should be fixed with the new major version and use of Aesara

Topic		Replies	Views
Understanding shape of values returned by sample_posterior_predictive Questions	3	915	January 26, 2020
Sample_ppc shape Questions	5	2687	July 18, 2018
Posterior predictive sampling with shared matrix Questions	2	596	August 31, 2018
Sample_prior_predictive bug, gives wrong shape for dependent variable? Questions	2	458	July 29, 2019
Shared theano in multiple regression Questions	12	1115	February 12, 2019