| Title: | Interface to the 'C++' Library 'Pf' |
|---|---|
| Description: | Builds and runs 'c++' code for classes that encapsulate state space model, particle filtering algorithm pairs. Algorithms include the Bootstrap Filter from Gordon et al. (1993) <doi:10.1049/ip-f-2.1993.0015>, the generic SISR filter, the Auxiliary Particle Filter from Pitt et al (1999) <doi:10.2307/2670179>, and a variety of Rao-Blackwellized particle filters inspired by Andrieu et al. (2002) <doi:10.1111/1467-9868.00363>. For more details on the 'c++' library 'pf', see Brown (2020) <doi:10.21105/joss.02599>. |
| Authors: | Taylor Brown [aut, cre]
|
| Maintainer: | Taylor Brown <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 1.0.1 |
| Built: | 2025-02-18 05:26:56 UTC |
| Source: | https://github.com/cran/pfr |
Build c++ particle filtering code for your R session.
buildModelFuncs(myDir, modelName, verbose = FALSE)buildModelFuncs(myDir, modelName, verbose = FALSE)
myDir |
directory with your three code files (i.e. model header, model source and export code) |
modelName |
your model name. Must be in all lowercase, and be a substring of the above-mentioned filenames |
verbose |
logical and passed in to inline::cxxfunction() |
an Rcpp Module object
## Not run: # compile everything from scratch svol_lev <- buildModelFuncs("~/Desktop", "svol_leverage") # then use your model's log-likelihood and filtering functions svol_lev$svol_leverage_bswc_approx_LL(rnorm(100), c(.9, 0.0, 1.0, -.2)) svol_lev$svol_leverage_bswc_approx_filt(rnorm(100), c(.9, 0.0, 1.0, -.2)) ## End(Not run)## Not run: # compile everything from scratch svol_lev <- buildModelFuncs("~/Desktop", "svol_leverage") # then use your model's log-likelihood and filtering functions svol_lev$svol_leverage_bswc_approx_LL(rnorm(100), c(.9, 0.0, 1.0, -.2)) svol_lev$svol_leverage_bswc_approx_filt(rnorm(100), c(.9, 0.0, 1.0, -.2)) ## End(Not run)
Create c++ template files for bootstrap filters (with or without covariates), auxiliary particle filters, sequential importance sampling with resampling filters, or Rao-Blackwellized/Marginal particle filters.
createPFCPPTemplates(modname, pfAlgo, fileDir, openNow = TRUE)createPFCPPTemplates(modname, pfAlgo, fileDir, openNow = TRUE)
modname |
name of model in all lowercase |
pfAlgo |
Either "BSF", "APF", "BSWC", "SISR", "RBPFHMM", or "RBPFKALMAN" |
fileDir |
where to save files. Not saved if NULL (but three files are returned in list). |
openNow |
TRUE if you want to open this now in RStudio. Ignored if fileDir is NULL. |
NULL if saving files, otherwise a list with three character vectors
# return in list of character strings createPFCPPTemplates("coolmod", "BSF", fileDir = NULL) ## Not run: # save three files to Desktop, and # begin editing them in rstudio IDE createPFCPPTemplates("coolmod", "BSF", fileDir = "~/Desktop/") ## End(Not run)# return in list of character strings createPFCPPTemplates("coolmod", "BSF", fileDir = NULL) ## Not run: # save three files to Desktop, and # begin editing them in rstudio IDE createPFCPPTemplates("coolmod", "BSF", fileDir = "~/Desktop/") ## End(Not run)