I am trying to build a model that seemingly consist of two normal distributions. My data however is binned. Therefore i believe a qaussian mixture model cannot be used(?). Is there a way that i can do a fitting of two normal distributions to binned data like this in pymc3? [image]

Fitting a histogram

junpenglao November 26, 2020, 2:59pm 7

you probably should adjust the prior - but if you have access to the binning it would be easier to fit a GMM directly, unless I am missing something here?

Topic		Replies	Views
Hierarchical model with mixture prior Questions	1	1801	January 12, 2019
Fitting multimodal data Questions	0	378	November 3, 2021
Fitting a spectra of gaussians Questions	6	3463	January 10, 2020
2D Gaussian Mixture Model Questions	7	2095	August 25, 2017
Truncated Distributions in Mixture Models? Questions	3	1355	September 20, 2017

Fitting a histogram

Related topics