Trace Plot

[Marimuthu]

Hi,

I ran 200,000 iterations with 10000 warmup and 100 thinning. Got the below trace plot. I can see there is a gap in the trace plot.

I got below warning messages:
Warning messages:
1: There were 29 divergent transitions after warmup. See
https://mc-stan.org/misc/warnings.html#divergent-transitions-after-warmup
to find out why this is a problem and how to eliminate them.
2: There were 4 chains where the estimated Bayesian Fraction of Missing Information was low. See
https://mc-stan.org/misc/warnings.html#bfmi-low
3: Examine the pairs() plot to diagnose sampling problems

Could anyone suggest can I ignore this or not? If not, please suggest a way to solve this. I appreciate your assistance.

Marimuthu

[js592]

In general, divergent transitions after warmup indicate problems with the HMC sampling and the results should’t be trusted. These problems can occasionally be the result of a genuinely difficult model to estimate but more often tend to indicate a need for re-parameterization and/or more careful priors.

It would be helpful if you could provide the code/model you are trying to estimate.

[Marimuthu]

Hi Jon,

I appreciate your prompt response. Let me try reparameterization.

Marimuthu.

[mitzimorris]

the Visualization in Bayesian Workflow paper has a section Moving beyond traceplots.

also, reparameterization - that’s a lot of iterations!

[Marimuthu]

Thank you, I will take a look at it.

Marimuthu