NIMBLE version 1.1.0 released

[Chris Paciorek]

We're pleased to announce the release of Version 1.1.0 of NIMBLE. This release provides new functionality as well as various bug fixes and improved error trapping, including:

• Improving our automatic differentiation (AD) system so it can be used in a wider range of models, including models with stochastic indexing, discrete latent states, and CAR distributions. Support for AD for these models means that HMC sampling and Laplace approximation can be used.
• Allowing distributions and functions (whether user-defined or built-in) that lack AD support (such as `dinterval`, `dconstraint`, and truncated distributions) to be used and compiled in AD-enabled models. The added flexibility increases the range of models in which one can use AD methods (HMC or Laplace) on some parts of a model and other samplers or methods on other parts.
• Adding `nimIntegrate` to the NIMBLE language, providing one-dimensional numerical integration via adaptive quadrature, equivalent to R's `integrate`. This can, for example, be used in a user-defined function or distribution for use in model code, such as to implement certain point process or survival models that involve a one-dimensional integral.
• Adding a "prior samples" MCMC sampler, which uses an existing set of numerical samples to define the prior distribution of model node(s).
• Better support of the dCRP distribution in non-standard model structures.
• Adding error trapping to prevent accidental use of  C++ keywords as model variable names.
• Removing the `RW_multinomial` MCMC sampler, which was found to generate incorrect posterior results (in cases when a latent state followed a multinomial distribution)
• Fixing a bug in conjugacy checking in a case of subsets of multivariate nodes.
• Fixing `is.na` and `is.nan` to operate in the expected vectorized fashion.
• Improving documentation of AD, nimbleHMC, and nimbleSMC in the manual.
• Updating Eigen (the C++ linear algebra library used by nimble) to version 3.4.0.

Please see the NEWS file in the installed package for more details.

- Chris Paciorek for the NIMBLE Development Team

[Edel Lheureux]

Hi Chris,

I am currently using a model with a multinomial distribution, but I encountered an error message stating that support for sampling multinomial distributions was removed from NIMBLE due to incorrect sampling results.

Is there a corrected version available, or would you recommend using a different sampler?

Thank you for your assistance.

Best regards,

Edel Lheureux

[Perry de Valpine]

Dear Edel,

There is not currently a sampler for latent multinomial models. We are aware of some interest in the user community in having this and would welcome a contribution. One of the challenges is that if N (size) itself is a latent variable that must be sampled by MCMC, it becomes necessary to sample it jointly with the multinomial outcome vector.

Best wishes,

Perry

--
You received this message because you are subscribed to the Google Groups "nimble-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nimble-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nimble-users/222a506f-2fd4-4f05-ae8f-3efc2184c787n%40googlegroups.com.

[Edel Lheureux]

Hi Perry, 

Thank you for your response, I would be interested to know if you found a solution and release a new version.

Best wishes, 

Edel Lheureux