Programming with Models: Statistical Algorithms for Hierarchical Model Structures using NIMBLE.

This IDSAI / Statistics co-hosted talk will introduce NIMBLE, a system for programming statistical algorithms for general hierarchical (graphical) model structures within R. NIMBLE is designed to meet three challenges: flexible model specification, a language for programming algorithms that can operate on these models, and a balance between high-level programmability and execution efficiency. For model specification, NIMBLE extends the BUGS language to define statistical models, which can manipulate variables, calculate log-density values, generate random simulations, and query the relationships among variables. For algorithmic programming, NIMBLE supports programmable functions that operate with model objects. To achieve efficient second-stage computation, NIMBLE compiles models and algorithms via C++. In addition to being a platform for programming, NIMBLE provides a pre-written library of statistical algorithms. These include both Markov chain Monte Carlo (MCMC) and sequential Monte Carlo (SMC) algorithms. We will describe these suites of algorithms, and demonstrate the power and flexibility of the NIMBLE system.