This repository gives the full reproducible code for generating all Figures and Tables in Sparse Projected Averaged Regression to High-dimensional Data (see Parzer, Filzmoser and Vana-Guer 2024).
Published version: Sparse data-driven random projection in regression for high-dimensional data, Parzer, Filzmoser and Vana-Guer 2025
This repository consists of the following folders with described contents.
- Chaper2Experiments: R-scripts and .rds-files for the experiments and Figures in chapter/section 2 'Methods'
- data: .txt files for the rat data and a .mat file for the face angel data
- data_application: R-script for data applications applying all methods on all data sets multiple times and saving the resulting .rds file to the folder 'saved_results'; and another R-script reproducing the preprocessing of the face data in 'Compressed Gaussian Process for Manifold Regression' (Guhaniyogi and Dunson 2016) for performance comparison
- functions: 4 R-scripts, data_generation.R for defining a function generating data from a HD linear model, methods.R defining consistent wrapper functions for all considered methods, multi_assign.R to define an operator assigning multiple variables at once (by Daniel Kapla, TU Wien) and RPM_generation.R to define three different functions generating certain random projection matrices (Sparse, Sparse_CW, our proposed adapted Sparse_CW)
- generate_plots: R-scipts reading in .rds files from 'saved_results' and generating the plots and tables for the simulation study and the data applications and saving the plots as pdfs in 'plots'
- plots: all pdf Figures
- saved_results: .rds files produced from 'simulation' or 'data_application' folders
- simulations: R-script for simulation study applying all methods on all simulation settings multiple times and saving the resulting .rds file to the folder 'saved_results'
- TARP-master: R-code from 'Targeted Random Projection for Prediction From High-Dimensional Features' (Mukhopadhyay and Dunson 2020) adapted to return the estimated beta regression coefficient
To enhance reproducibility, the following output was obtained from sessionInfo() after executing the header of the file 'simulations_SPAR_CV_022024.R'.
R version 4.2.1 (2022-06-23)
Platform: aarch64-apple-darwin20 (64-bit)
Running under: macOS 14.3.1
Matrix products: default
LAPACK: /Library/Frameworks/R.framework/Versions/4.2-arm64/Resources/lib/libRlapack.dylib
locale:
[1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
attached base packages:
[1] parallel stats graphics grDevices utils datasets methods base
other attached packages:
[1] spar_4.0.0 stringr_1.5.1 robustHD_0.8.0 robustbase_0.99-0 perry_0.3.1 ggplot2_3.5.1 SplitReg_1.0.2 MASS_7.3-60
[9] SIS_0.8-8 glmnet_4.1-8 pls_2.8-1 Matrix_1.6-1.1 ROCR_1.0-11 dplyr_1.1.4 tidyr_1.3.1 foreach_1.5.2
loaded via a namespace (and not attached):
[1] Rcpp_1.0.13-1 DEoptimR_1.0-11 pillar_1.9.0 compiler_4.2.1 iterators_1.0.14 tools_4.2.1 lifecycle_1.0.4 tibble_3.2.1 gtable_0.3.6
[10] lattice_0.20-45 pkgconfig_2.0.3 rlang_1.1.4 cli_3.6.3 rstudioapi_0.13 withr_3.0.2 generics_0.1.3 vctrs_0.6.5 grid_4.2.1
[19] tidyselect_1.2.1 glue_1.8.0 R6_2.5.1 fansi_1.0.6 Rdpack_2.6.2 survival_3.3-1 pacman_0.5.1 purrr_1.0.2 magrittr_2.0.3
[28] rbibutils_2.3 scales_1.3.0 codetools_0.2-18 splines_4.2.1 colorspace_2.1-1 shape_1.4.6.1 ncvreg_3.14.1 utf8_1.2.4 stringi_1.8.4
[37] munsell_0.5.1