Hi @ricardoV94
Thanks for responding. I have little to no experience in PYMC and I am trying to get into deep waters here. I am trying to implement a cox proportional hazards model for competing risks.
I have specified the model and it is sampling. I am estimating mean and standard deviation of the baseline hazard rate and regression parameters in proportional hazards. My problem looks like this
I have a custom likelihood that I have implemented.
While sampling I can see divergences. Also the parameters mean and standard deviation have almost constant values in the generated samples in respective chains, so the chains look stuck.
I have a dataset which has approx 1.2M records spanning from 2013-19. I have random sampled it down to 60k for testing the model.
For diagnostics, I dont have a clue where to start. @ricardoV94 if you could suggest something?
