Help with Multivariate SDE Timeseries

[Marion2025]

@junpenglao

Oh yes I remember @twiecki mentioned somewhere that if you have double stochastic (i.e., random walk of variance plus random walk of mean) it is extremely difficult to do inference.

I don’t think this particular problem is that difficult, it can be done using Metropolis but it tends to be slow. I think NUTS can handle it if we specify the model correctly.