SCPANOVA

SPCANOVA is a GitHub repository containing all the figures and simulations results shown in the paper

@article{NAVARROGARCIA2024,
title = {On mathematical optimization for shape-constrained generalized additive models using $P-$splines},
year = {2024},
url = {https://www.researchgate.net/publication/370492139},
author = {Navarro-García, Manuel and Guerrero, Vanesa and Durbán, María},
keywords = {Shape-constrained regression, Generalized additive models, Penalized splines, Conic optimization},
}

All the simulation studies carried out in this work use the routines implemented in cpsplines, which requires a MOSEK license to solve the optimization problems.

Project structure

The current version of the project is structured as follows:

• data: a folder containing CSV and parquet files with simulated and real data sets, and the results of the applications.
• scpanova: the main directory of the project, which consist of:
  • additive_model.py: contains the code to fit constrained additive P-splines.
  • anova_model.py: contains the code to fit constrained ANOVA P-splines.
  • figures.ipynb: A Jupyter notebook containing the code used to generate the figures and the tables of the paper.
  • hschool.py: the code used to carry out the case-study with the hschool data set.
  • pima.py: the code used to carry out the case-study with the pima data set.
  • production_function.R: contains the R code of the package AAFS used during the real case-studies.
  • R_scripts.R: contains the R code of the packages scam and cgam used during the real case-studies.
  • R_utils.R: an R script with some useful functions.
  • simulations_production.py: the code used to carry out the simulation study for estimating production functions.
  • simulations.py: the code used to carry out the simulation study.

Package dependencies

generalized_cpsplines_multi mainly depends on the following packages:

• cpsplines.
• pyarrow
• rpy2
• scikit-learn
• typer

Contact Information and Citation

If you have encountered any problem or doubt while using SPCANOVA, please feel free to let me know by sending me an email:

• Name: Manuel Navarro García (he/his)
• Email: manuelnavarrogithub@gmail.com

If you find SPCANOVA or cpsplines useful, please cite it in your publications.