Hi there, I though it was a simple matter but I’m getting nowhere with this. I’m basically “just” trying to construct an asymmetric StudentT distribution, i.e., two distributions with different sigmas, hoping to find a way to treat them as one and/or to normalize them so their peak value is the sam…

The reason I asked what the modeled is doing is because that might change how to approach your problem. mu = pm.Normal('mu', 0, 1) sigmas = pm.HalfNormal('sigmas', sigma=[1, 2], shape=2) # You will probably need to use an ordered transform to avoid symmetries sigma = tt.switch(tt.gt(observed, mode…

Combining two distributions

Questions

ricardoV94 January 25, 2021, 8:29pm 2

What is the “modeled” variable in your example doing?

Topic		Replies	Views
How to create Non-Central Student's T distribution and what priors to use with the distribution? Questions	10	1941	June 14, 2020
"Non-Central" Student T Questions	4	825	October 30, 2017
Mixture of Multivariate Student-T Questions	1	812	November 7, 2018
How to force estimated variables into another distribution variable Questions	6	667	January 9, 2020
pm.HalfStudentT.dist() failing to give correct shape v5 bug	6	379	April 7, 2023

Combining two distributions

Related topics