Just a quick comment. Joint mode and funnel can even have the opposite behavior: as I showed in my simulation, in models quite like “Gelman’s funnel”, increasing group variance eventually eliminates funnel as tau is kept away from 0 but also leads to bimodality, in contrast decreasing group variance will avoid bimodality but gives rise to the funnel/metastability/slow mixing of HMC chains.
yuling
8
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| EBFMI (very) low, looking for reparameterization advice | 27 | 2140 | October 1, 2021 | |
| Prior recommendation for scale parameters in hierarchical models too strong? | 25 | 8964 | January 31, 2018 | |
| Centered vs non-centered model, different loglikelihoods | 6 | 955 | May 15, 2024 | |
| Adaptation and non-centered parameterisations | 10 | 1283 | August 2, 2021 | |
| Zaps to avoid the funnel | 9 | 1737 | October 3, 2018 |