Routines for constructing sparse matrices of basis functions, as well as a custom implementation of L1-regularized regression (i.e. the LASSO), altogether providing a fast, computationally efficient implementation of the Highly Adaptive LASSO (HAL), primarily in C++ but made available for use in R through Rcpp.

hal9001()

References

Documentation:

https://tlverse.org/hal9001

GitHub repository:

https://github.com/tlverse/hal9001