Tools for performing inference and analysis on a class of quadratic diffusion processes for both scalar and bivariate diffusion systems. For scalar diffusions, a module is provided for solving first passage time problems for both time-homogeneous and time-inhomogeneous GQDs.
Inference and Analysis for Generalized Quadratic Diffusions
DiffusionRgqd is collection of tools for performing inference and analysis on scalar and bivariate time-inhomogeneous diffusion processes with quadratic drift and diffusion terms in R.
DiffusionRgqd provides a simple interface that requires minimal mathematical input in order to perform analysis on non-linear, time-inhomogeneous diffusion processes. The package also makes use of C++ in order to maximize the computational efficiency of inference routines. As such it is possible to conduct inference on a plethora of models in a desktop environment without incurring excessively long computation times.
Mac users may have to carry out some additional installation procedures in order for DiffusionRjgqd to operate optimally.
To install the latest version of Rcpp, the latest version of R is needed. To install RcppArmadillo, the Fortran version used by R needs to be updated. To install rgl, the computer needs to have X11 installed. Update R to the latest version. Run the following code: install.packages("Rcpp", type = "source",dep=T)
curl -O http://r.research.att.com/libs/gfortran-4.8.2-darwin13.tar.bz2 sudo tar fvxz gfortran-4.8.2-darwin13.tar.bz2 -C /
Go to Applications/Utilities and see if X11 is there. If not, you’ll need to install X11 or XQuartz. These are available from http://xquartz.macosforge.org/landing/
install.packages("~/DiffusionRgqd_0.1.3.tar.gz", repos = NULL, type = "source”)
library(DiffpackRgqd) example(GQD.density) example(GQD.mcmc)
DiffusionRgqd is now available on CRAN and GitHub.