Mauricio Vargas Sepulveda
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Mauricio Vargas Sepulveda has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
Conrad Sanderson
2 shared publications- Jeroen Ooms
Berkeley Institute for Data Science
2 shared publications Joshua Kunst
Pontifical Catholic University of Chile
2 shared publicationsLital Barkai
Independent Software Developer
2 shared publicationsPbc Posit Software
2 shared publicationsAnna-Lena Woelwer
1 shared publication
Latest Publications
- 3 September 2026
- Published by The R Foundation
later2: Utilities for Scheduling Functions to Execute Later with Event Loops
Executes arbitrary R or C functions some time after the current time, after the R execution stack has emptied. The functions are scheduled in an event loop. This is a derived work from the 'later' package aiming to reduce the number of dependencies.
- 3 September 2026
- Published by The R Foundation
later2: Utilities for Scheduling Functions to Execute Later with Event Loops
Executes arbitrary R or C functions some time after the current time, after the R execution stack has emptied. The functions are scheduled in an event loop. This is a derived work from the 'later' package aiming to reduce the number of dependencies.
- Open Access
- 29 August 2026
- Published by Elsevier in SSRN Electronic Journal
Capybara 2.0: Generalized Linear Models with High-dimensional Fixed Effects
This paper presents a new version of capybara, an R package implementing computationally efficient algorithms for estimating generalized linear models (GLMs) with high-dimensional fixed effects. Building on Stammann (2018) and Berge (2018), we provide an alternative approach to full design matrix with fixed effects that is faster and reduces memory footprint without sacrificing numerical accuracy. The package is particularly valuable for trade economics or any application requiring multiple high-dimensional fixed effects to control for unobserved heterogeneity, making previously infeasible models computationally tractable on standard consumer-level hardware.
- 29 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessCapybara 2.0: Generalized Linear Models with High-dimensional Fixed Effects
This paper presents a new version of capybara, an R package implementing computationally efficient algorithms for estimating generalized linear models (GLMs) with high-dimensional fixed effects. Building on Stammann (2018) and Berge (2018), we provide an alternative approach to full design matrix with fixed effects that is faster and reduces memory footprint without sacrificing numerical accuracy. The package is particularly valuable for trade economics or any application requiring multiple high-dimensional fixed effects to control for unobserved heterogeneity, making previously infeasible models computationally tractable on standard consumer-level hardware.
- 7 August 2026
- Published by The R Foundation
tinycache: Cache Objects in Disk or Memory
Reuse objects with a long processing time either by storing those in disk or memory. Uses caching to identify R objects (e.g., data frames, plots, etc.) and allows repeated access to those. Created with the specific goal of skipping the waiting time for summary tables obtained from large 'SQL' tables. It is extensible to other uses, such as caching plots in 'Tabler' dashboards to reduce waiting times.
- 7 August 2026
- Published by The R Foundation
tinycache: Cache Objects in Disk or Memory
Reuse objects with a long processing time either by storing those in disk or memory. Uses caching to identify R objects (e.g., data frames, plots, etc.) and allows repeated access to those. Created with the specific goal of skipping the waiting time for summary tables obtained from large 'SQL' tables. It is extensible to other uses, such as caching plots in 'Tabler' dashboards to reduce waiting times.
- 6 August 2026
- Published by The R Foundation
tinyroxygen: A Tiny 'Roxygen'-Style Documentation Generator
A minimalistic, dependency-free alternative to 'roxygen2'. Reads '#” comment blocks above R functions and objects and turns them into '.Rd' documentation files and a 'NAMESPACE' file. 'tinyroxygen' is to 'roxygen2' what 'tinytest' is to 'testthat', a lightweight alternative built with base R only, with no recursive dependencies to install. 'Roxygen' stands for R documentation inspired by the 'Doxygen' for 'C++' documentation tools.
- 6 August 2026
- Published by The R Foundation
tinyroxygen: A Tiny 'Roxygen'-Style Documentation Generator
A minimalistic, dependency-free alternative to 'roxygen2'. Reads '#” comment blocks above R functions and objects and turns them into '.Rd' documentation files and a 'NAMESPACE' file. 'tinyroxygen' is to 'roxygen2' what 'tinytest' is to 'testthat', a lightweight alternative built with base R only, with no recursive dependencies to install. 'Roxygen' stands for R documentation inspired by the 'Doxygen' for 'C++' documentation tools.
- Open Access
- 10 March 2026
- Published by Elsevier in SSRN Electronic Journal
A Memory-Efficient Rank-Revealing Cholesky Decomposition Algorithm
We present a rank-revealing Cholesky decomposition algorithm for symmetric positive semi-definite matrices, designed to identify and handle linear dependencies efficiently. The proposed method iteratively computes the Cholesky factorization while checking for near-zero diagonal entries, allowing for the detection of rank deficiencies based on a user-defined tolerance. Implemented using the Armadillo C++ library, this approach leverages optimized linear algebra routines to ensure computational efficiency. Benchmarks against matrix decomposition methods demonstrate the algorithm's effectiveness in practical applications, particularly in the context of Generalized Linear Models (GLMs) where multicollinearity often introduces estimation challenges.
- 10 March 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Memory-Efficient Rank-Revealing Cholesky Decomposition Algorithm
We present a rank-revealing Cholesky decomposition algorithm for symmetric positive semi-definite matrices, designed to identify and handle linear dependencies efficiently. The proposed method iteratively computes the Cholesky factorization while checking for near-zero diagonal entries, allowing for the detection of rank deficiencies based on a user-defined tolerance. Implemented using the Armadillo C++ library, this approach leverages optimized linear algebra routines to ensure computational efficiency. Benchmarks against matrix decomposition methods demonstrate the algorithm's effectiveness in practical applications, particularly in the context of Generalized Linear Models (GLMs) where multicollinearity often introduces estimation challenges.
- 3 November 2025
- Published by The R Foundation
armadillo4r: An 'Armadillo' Interface
Provides function declarations and inline function definitions that facilitate communication between R and the 'Armadillo' 'C++' library for linear algebra and scientific computing. This implementation is derived from Vargas Sepulveda and Schneider Malamud (2024) .
- 3 November 2025
- Published by The R Foundation
armadillo4r: An 'Armadillo' Interface
Provides function declarations and inline function definitions that facilitate communication between R and the 'Armadillo' 'C++' library for linear algebra and scientific computing. This implementation is derived from Vargas Sepulveda and Schneider Malamud (2024) .
- 16 October 2025
- Published by The R Foundation
cpp4r: Header-Only 'C++' and 'R' Interface
Provides a header only, 'C++' interface to 'R' with enhancements over 'cpp11'. Enforces copy-on-write semantics consistent with 'R' behavior. Offers native support for ALTREP objects, 'UTF-8' string handling, modern 'C++11' features and idioms, and reduced memory requirements. Allows for vendoring, making it useful for restricted environments. Compared to 'cpp11', it adds support for converting 'C++' maps to 'R' lists, 'Roxygen' documentation directly in 'C++' code, proper handling of matrix attributes, support for nullable external pointers, bidirectional copy of complex number types, flexibility in type conversions, use of nullable pointers, and various performance optimizations.
- 16 October 2025
- Published by The R Foundation
cpp4r: Header-Only 'C++' and 'R' Interface
Provides a header only, 'C++' interface to 'R' with enhancements over 'cpp11'. Enforces copy-on-write semantics consistent with 'R' behavior. Offers native support for ALTREP objects, 'UTF-8' string handling, modern 'C++11' features and idioms, and reduced memory requirements. Allows for vendoring, making it useful for restricted environments. Compared to 'cpp11', it adds support for converting 'C++' maps to 'R' lists, 'Roxygen' documentation directly in 'C++' code, proper handling of matrix attributes, support for nullable external pointers, bidirectional copy of complex number types, flexibility in type conversions, use of nullable pointers, and various performance optimizations.
- Open Access
- 27 September 2025
- Published by The R Foundation
spuriouscorrelations: Datasets with Strong and Spurious Correlations
Provides datasets from Vigen (2015) rescued from the Internet Wayback Machine. These should be preserved for statistics introductory courses as these make it very clear that correlation is not causation.
- 27 September 2025
- Published by The R Foundation
Open Accessspuriouscorrelations: Datasets with Strong and Spurious Correlations
Provides datasets from Vigen (2015) rescued from the Internet Wayback Machine. These should be preserved for statistics introductory courses as these make it very clear that correlation is not causation.
- Open Access
- 16 September 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (9) , e0331178
Capybara: Efficient estimation of generalized linear models with high-dimensional fixed effects
This paper introduces capybara, an R package implementing computationally efficient algorithms for estimating generalized linear models (GLMs) with high-dimensional fixed effects. Building on Stammann (2018), we combine the Frisch-Waugh-Lovell (FWL) theorem with alternating projections to achieve memory-efficient estimation. Our benchmarks demonstrate that capybara reduces computation time by 95-99% compared to traditional dummy variable approaches while maintaining numerical accuracy to 5 decimal places. For a complex gravity model with 28,000 observations and 3,200 fixed effects, capybara completes estimation in just 6 seconds using 33 MB of memory, compared to 11 minutes and 12 GB with base R. The package is particularly valuable for trade economics, labor economics, and other applications requiring multiple high-dimensional fixed effects to control for unobserved heterogeneity, making previously infeasible models computationally tractable on standard hardware.
- 16 September 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (9) , e0331178
Open AccessCapybara: Efficient estimation of generalized linear models with high-dimensional fixed effects
This paper introduces capybara, an R package implementing computationally efficient algorithms for estimating generalized linear models (GLMs) with high-dimensional fixed effects. Building on Stammann (2018), we combine the Frisch-Waugh-Lovell (FWL) theorem with alternating projections to achieve memory-efficient estimation. Our benchmarks demonstrate that capybara reduces computation time by 95-99% compared to traditional dummy variable approaches while maintaining numerical accuracy to 5 decimal places. For a complex gravity model with 28,000 observations and 3,200 fixed effects, capybara completes estimation in just 6 seconds using 33 MB of memory, compared to 11 minutes and 12 GB with base R. The package is particularly valuable for trade economics, labor economics, and other applications requiring multiple high-dimensional fixed effects to control for unobserved heterogeneity, making previously infeasible models computationally tractable on standard hardware.