Skip to main content

Research Repository

Advanced Search

Variable selection for inferential models with relatively high-dimensional data: Between method heterogeneity and covariate stability as adjuncts to robust selection

Lima, Eliana; Davies, Peers; Kaler, Jasmeet; Lovatt, Fiona; Green, Martin

Authors

Eliana Lima

Peers Davies

JASMEET KALER JASMEET.KALER@NOTTINGHAM.AC.UK
Professor of Epidemiology & Precision Livestock Informatics

FIONA LOVATT FIONA.LOVATT@NOTTINGHAM.AC.UK
Clinical Associate Professor

MARTIN GREEN martin.green@nottingham.ac.uk
Professor of Cattle Health & Epidemiology



Abstract

Variable selection in inferential modelling is problematic when the number of variables is large relative to the number of data points, especially when multicollinearity is present. A variety of techniques have been described to identify ‘important’ subsets of variables from within a large parameter space but these may produce different results which creates difficulties with inference and reproducibility. Our aim was evaluate the extent to which variable selection would change depending on statistical approach and whether triangulation across methods could enhance data interpretation. A real dataset containing 408 subjects, 337 explanatory variables and a normally distributed outcome was used. We show that with model hyperparameters optimised to minimise cross validation error, ten methods of automated variable selection produced markedly different results; different variables were selected and model sparsity varied greatly. Comparison between multiple methods provided valuable additional insights. Two variables that were consistently selected and stable across all methods accounted for the majority of the explainable variability; these were the most plausible important candidate variables. Further variables of importance were identified from evaluating selection stability across all methods. In conclusion, triangulation of results across methods, including use of covariate stability, can greatly enhance data interpretation and confidence in variable selection.

Journal Article Type Article
Publication Date May 14, 2020
Journal Scientific Reports
Print ISSN 2045-2322
Electronic ISSN 2045-2322
Publisher Nature Publishing Group
Peer Reviewed Peer Reviewed
Volume 10
Article Number 8002
APA6 Citation Lima, E., Davies, P., Kaler, J., Lovatt, F., & Green, M. (2020). Variable selection for inferential models with relatively high-dimensional data: Between method heterogeneity and covariate stability as adjuncts to robust selection. Scientific Reports, 10, https://doi.org/10.1038/s41598-020-64829-0
DOI https://doi.org/10.1038/s41598-020-64829-0
Keywords Multidisciplinary
Publisher URL https://www.nature.com/articles/s41598-020-64829-0
Additional Information Received: 29 January 2020; Accepted: 17 April 2020; First Online: 14 May 2020; : The authors declare no competing interests.

Files

Variable selection for inferential models with relatively highdimensional data: Between method heterogeneity and covariate stability as adjuncts to robust selection (1.6 Mb)
PDF





You might also like



Downloadable Citations

;