How to use these vignettes

Notation

In the textbook, the notation is of the form:

$$Y_i = \beta_1 X_{i1} + \beta_2 X_{i2} + \ldots + \beta_k X_{ik} + \beta_0 + e_i,$$

where $Y_i$ is the dependent variable, $X_{ij}$ are the independent variables, $\beta_j$ are the coefficients, and $e_i$ is the error term.

In these notes I use the following notation:

$$Y_i = \beta_0 + \beta_1 X_{i1} + \beta_2 X_{i2} + \ldots + \beta_k X_{ik} + e_i.$$

Code structure

For the code, I use two functions, one with the suffix _ that receives the data in C++ data types and another without the suffix that receives the data in R data types. The function with the suffix _ is the one that does the actual work, and the function without the suffix is the one that converts the data from R to C++ data types and vice versa.

[C++ data type] my_function_([C++ data type] a, [C++ data type] b) {
  ... C++ code ...

  return result;
}

[[cpp11::register]] [R data type] my_function_r_([R data type] a,
                                                 [R data type] b) {
  [C++ data type] A = as_[C++ data type](a);
  [C++ data type] B = as_[C++ data type](b);

  [C++ data type] result = my_function_(A, B);

  return as_[R data type](result);
}

This allows to reuse functions in the C++ code more easily and also avoids overhead coping data between R and C++ multiple times.

The functions must be exposed and documented so that these can be used by the end-user:

#' Add Title
#' @param a Description
#' @param b Description
#' @return matrix, vector, etc containing the result
#' @export
my_function <- function(a, b) {
  my_function_r_(a, b)
}

Please look at the code in the ‘R’ and ‘src’ folders in the package repository.

The code uses a header used across all the scripts in order to avoid importing the libraries multiple times and it allows us to use the ols_ function across all the scripts. The header is:

#include <cpp11.hpp>
#include <cpp11armadillo.hpp>
#include <unordered_map>

using namespace arma;
using namespace cpp11;

Mat<double> ols_(const Mat<double>& Y, const Mat<double>& X);