pomp (version 0.39-3)

pomp-package: Partially-observed Markov processes

Description

The pomp package provides facilities for inference on time series data using partially-observed Markov processes (AKA state-space models or nonlinear stochastic dynamical systems). One can use pomp to fit nonlinear, non-Gaussian dynamic models to time-series data. The first step in using pomp is to encode one's model and data in an object of class pomp. One does this via a call to pomp, which involves specifying the process and measurement components of the model in one or more of a variety of ways. Details on this are given in the documentation for the pomp function and examples are given in the intro_to_pomp vignette.

Currently, pomp provides algorithms for (i) simulation of stochastic dynamical systems (see simulate), (ii) particle filtering (AKA sequential Monte Carlo or sequential importance sampling), see pfilter), (iii) the iterated filtering method of Ionides et al. (2006), see mif), (iv) the nonlinear forecasting algorithm of Kendall et al. (2005), see nlf), (v) the particle MCMC approach of Andrieu et al. (2010), see pmcmc, (vi) basic trajectory matching, see traj.match, (vi) the probe-matching method of Wood (2010) and Kendall et al. (1999), see probe.match, (vii) a spectral probe-matching method (Reuman et al., 2006), see spect.match. See the package website http://pomp.r-forge.r-project.org for these references. The package also provides various tools for plotting and extracting information on models and data as well as an API for algorithm development. Future support for additional algorithms in envisioned, and implementations of the Bayesian sequential Monte Carlo approach of Liu & West. Much of the work in pomp has been done under the auspices of a working group of the National Center for Ecological Analysis and Synthesis (NCEAS), "Inference for Mechanistic Models".

The package is provided under the GNU Public License (GPL). Contributions are welcome, as are comments, suggestions for improvements, and bug reports. See the package website http://pomp.r-forge.r-project.org for more information, access to the package mailing list, links to the authors' websites, and references to the literature.

Classes

pomp makes extensive use of S4 classes. The basic class, pomp, encodes a partially-observed Markov process together with a uni- or multi-variate data set and (possibly) parameters.

Vignettes

The vignette Introduction to pomp illustrates the facilities of the package using familiar stochastic processes. Run vignette("intro_to_pomp") or look at the HTML documentation to view the vignette. Methods for accelerating your codes are discussed in the Advanced topics in pomp vignette; run vignette("advanced_topics_in_pomp") to view it.

See Also

pomp, pfilter, simulate, trajectory, mif, nlf, probe.match, traj.match, bsmc, pmcmc