# Matthew Fidler

#### 6 packages on CRAN

A Scannerless GLR parser/parser generator. Note that GLR standing for "generalized LR", where L stands for "left-to-right" and R stands for "rightmost (derivation)". For more information see <https://en.wikipedia.org/wiki/GLR_parser>. This parser is based on the Tomita (1987) algorithm. (Paper can be found at <http://acl-arc.comp.nus.edu.sg/archives/acl-arc-090501d3/data/pdf/anthology-PDF/J/J87/J87-1004.pdf>). The original 'dparser' package documentation can be found at <http://dparser.sourceforge.net/>. This allows you to add mini-languages to R (like RxODE's ODE mini-language Wang, Hallow, and James 2015 <DOI:10.1002/psp4.12052>) or to parse other languages like 'NONMEM' to automatically translate them to R code. To use this in your code, add a LinkingTo dparser in your DESCRIPTION file and instead of using #include <dparse.h> use #include <dparser.h>. This also provides a R-based port of the make_dparser <http://dparser.sourceforge.net/d/make_dparser.cat> command called mkdparser(). Additionally you can parse an arbitrary grammar within R using the dparse() function, which works on most OSes and is mainly for grammar testing. The fastest parsing, of course, occurs at the C level, and is suggested.

Provides a simple mechanism to specify a symmetric block diagonal matrices (often used for covariance matrices). This is based on the domain specific language implemented in 'nlmixr' but expanded to create matrices in R generally instead of specifying parts of matrices to estimate.

Provides 'Scilab' 'n1qn1', or Quasi-Newton BFGS "qn" without constraints and 'qnbd' or Quasi-Newton BFGS with constraints. This takes more memory than traditional L-BFGS. The n1qn1 routine is useful since it allows prespecification of a Hessian. If the Hessian is near enough the truth in optimization it can speed up the optimization problem. Both algorithms are described in the 'Scilab' optimization documentation located at <http://www.scilab.org/content/download/250/1714/file/optimization_in_scilab.pdf>.

Most of the time floating point arithmetic does approximately the right thing. When adding sums or having products of numbers that greatly differ in magnitude, the floating point arithmetic may be incorrect. This package implements the Kahan (1965) sum <doi:10.1145/363707.363723>, Neumaier (1974) sum <doi:10.1002/zamm.19740540106>, pairwise-sum (adapted from 'NumPy', See Castaldo (2008) <doi:10.1137/070679946> for a discussion of accuracy), and arbitrary precision sum (adapted from the fsum in 'Python' ; Shewchuk (1997) <http://www.cs.berkeley.edu/~jrs/papers/robustr.pdf>). In addition, products are changed to long double precision for accuracy, or changed into a log-sum for accuracy.

Fit and compare nonlinear mixed-effects models in differential equations with flexible dosing information commonly seen in pharmacokinetics and pharmacodynamics (Almquist, Leander, and Jirstrand 2015 <doi:10.1007/s10928-015-9409-1>). Differential equation solving is by compiled C code provided in the 'RxODE' package (Wang, Hallow, and James 2015 <doi:10.1002/psp4.12052>).

Facilities for running simulations from ordinary differential equation (ODE) models, such as pharmacometrics and other compartmental models. A compilation manager translates the ODE model into C, compiles it, and dynamically loads the object code into R for improved computational efficiency. An event table object facilitates the specification of complex dosing regimens (optional) and sampling schedules. NB: The use of this package requires both C and Fortran compilers, for details on their use with R please see Section 6.3, Appendix A, and Appendix D in the "R Administration and Installation" manual. Also the code is mostly released under GPL. The VODE and LSODA are in the public domain. The information is available in the inst/COPYRIGHTS.