There are no relevant documents available to address your query about sampling options with many observations and a HurdleGamma likelihood. However, based on general knowledge, when dealing with divergences in sampling, especially with complex likelihoods like HurdleGamma, you might consider the following approaches:
-
Reparameterization: Sometimes, reparameterizing the model can help reduce divergences. This involves changing the parameterization of your model to make it more stable.
-
Tuning: Increase the number of tuning steps. This allows the sampler more time to adapt to the posterior distribution.
-
Step Size Adjustment: Manually adjusting the step size can sometimes help. A smaller step size can reduce divergences but may increase computation time.
-
Prior Specification: Ensure that the priors are well-specified and informative enough to guide the sampler, especially if you have not set a prior for the psi parameter.
-
Alternative Samplers: If numpyro and blackjax are not working, consider trying other samplers or adjusting their settings.
-
Data Subsetting: If possible, continue using a smaller subset of data to diagnose and fix issues before scaling up.
These strategies can help in addressing sampling issues with complex likelihoods.
For more information, visit https://www.pymc.io/
This answer was generated by the PyMC Assistant bot in response to your command. Please let us know if it was helpful!