
#14 Hidden Markov Models & Statistical Ecology, with Vianey Leos-Barajas
Learning Bayesian Statistics
00:00
How to Estimate Unidentifiable Parameters in Your Models?
It's a very common problem to have unidentifiable parameters in your models. And so usually, as I said, one of the consequences is that your chains won't even sample correctly or won't converge. So for me, it really has to be on a case by case basis. When I see my posterior distribution doing very strange things, when I end up with extremely multi-modal posterior distributions, I start to realize like, okay, something is off.
Transcript
Play full episode