I have a model where there is this mixture-esque deterministic equation (p * exp(-k1*d)) + ((1-p) * exp(-k2*d)) where p, k1, and k2 are stochastic variables and d is data. The samples being returned are seemingly sensible, it’s just there’s no chain convergence. For example, samples of chain 1 of k…

Yes, you are seeing label switching in your trace. It happens a lot with mixture models. One of the must-reads related to this topic is Michael Betancourt’s Stan case study Identifying Bayesian Mixture Models . I have a Gist here replicating all the experiment in PyMC3. As an experiment, you can al…

Constrain one variable to be greater than another

Questions

drbenvincent January 29, 2018, 9:06pm 13

Bler. I think I’m going to need to sketch that out in order to follow that explanation. Unless someone wants to write a blog post with animated explanatory diagrams

Topic		Replies	Views
Constraining order in a Mixture model Questions	6	1440	August 9, 2021
Mixture models and breaking class symmetry Questions	11	3932	October 12, 2018
Why does `transform=pm.distributions.transforms.ordered` lead to worse convergence? Questions	3	1579	August 27, 2021
Mixture of bounded normals as a prior Questions	4	1363	April 27, 2018
Error using Ordered() in Mixture of Uniform Distributions Questions	14	1899	February 7, 2019

Constrain one variable to be greater than another

Related topics