I have a dataset that has a bimodal distribution, and I want to understand what influences the chance of a data point being in one mode or the other. Here’s my model: # classes = 2, predictors.shape = (129,9) and data.shape = (129,) with pm.Model(coords={"classes": range(classes), "predictors": pre…

I figured it out. Changing the w term to this fixes the issue: w = pm.math.stack([p, 1-p], axis=1) Thanks!

Problem with NormalMixture

twiecki August 31, 2023, 12:58pm 6

Thanks for posting the solution!

Topic		Replies	Views
Problem with fitting Multivariate Mixture of Gaussians Questions	6	3561	June 4, 2018
Dirichlet process mixtures bidimensional data Questions	16	1229	February 13, 2019
Shape confusion in discrete mixture Questions	0	433	October 15, 2020
Model multivariate normal with separate means, dimension mismatch error Questions	2	3090	August 7, 2017
Help Debugging Dirichlet Process Mixture of Multivariate Normal Questions	0	536	February 17, 2021