My students and I are trying to use PyMC to explore the parameter space that could produce a particular data set (in astrophysics, if it matters.) One complication is that each observed data point can actually rule out a region of parameter space, i.e., \ln(\ell(\vec{\Theta})) \to - \infty for cert…

Can Hamiltonian MCMC be used effectively if the likelihood is identically zero for some parameters?

ricardoV94 June 10, 2025, 8:22pm 2

Can you tell a bit more about the structure of the likelihood?

Topic		Replies	Views
Limit range of parameters in hamiltonian monte carlo Questions	5	1379	May 5, 2019
Derivative warning while using Hamiltonian Monte Carlo Questions	1	578	May 17, 2019
UserWarning after sampling Questions	9	7105	July 17, 2017
Reject invalid parameter vector during HMC sampling Questions	2	483	April 16, 2020
OverflowError: math range error Questions	4	3656	March 18, 2018