Last updated: 26rd February 2015
- Create the code for the simulation of the marked point process in C++. Ox's idiom for returning multiple objects by making them global and using functions to fill them up is nasty.
- Objects from this function are to be returned as a list.
- Create a new file for the R version of the marked point process code, before completing the C++ version.
- Recall that for Ox, even vectors are matrices, and exactly the other way around for R -- even matrices are vectors.
- Completed the R version of the marked point process simulation code.
- Pesky issue with
useDynLibnot being recognized byroxygen2. Need to add aNULLto the end of the package file to get it to do this. - The
CrealKoopmanLucas2013.Rexample file should now work. - Remove the
Rversion of the Gaussian GAS copula model. - Remove the
Rversion of the function for estimating restricted values of the parameters.
-
Create the function that optimizes the GAS Gaussian copula model as in Creal, Koopman & Lucas (2013).
-
TODO: delete the R versions of the GAS Gaussian copula likelihood functions, or make them loadable when requested with the
Ronlyswitch. This is intended to be similar in functionality to astataonlyswitch that switches off Mata in Stata. -
Create the function that optimizes the marked point process model as in Creal, Koopman & Lucas (2013).