Hi everyone, I need to fit some experimental data using a model that can be expressed as: y(t)=N(S_0.f(t, \theta),\sigma^2 where S_0 is a scaling parameter, f(t, \theta) is a non linear function of the time t and of the model parameters \theta, and \sigma is the standard deviation of the noise. Us…

The error would be improve in the next release, but currently you need to do some manual check by hand when you see [image] tboutelier: ValueError: Bad initial energy: nan. The model might be misspecified. For example, here when you do In [12]: model.check_test_point() Out[12]: theta …

Custom likelihood function (again) with DensityDist and Theano function

Questions

junpenglao November 6, 2018, 9:22am 4

I always use switch, but I havent test the differences personally so I dont know as well.

Topic		Replies	Views
How to set up a custom likelihood function for two variables Questions	8	13306	March 9, 2018
Defining a numeric (custom) likelihood function in PyMC3 Questions theano	2	3070	August 25, 2018
How to do linear fitting with pymc3 by using pm.DensityDist Questions	6	1046	May 4, 2018
Using non-Theano function in likelihood, complicated model for optimal experimental design Questions	0	402	May 22, 2020
Fitting Custom Distributions and Likelihoods Questions theano	5	1135	July 14, 2021

Custom likelihood function (again) with DensityDist and Theano function

Related topics