Variable selection in social-environmental data: Sparse regression and tree ensemble machine learning approaches

TL;DR

This study evaluates machine learning methods for selecting social-environmental variables from census data that are truly associated with health outcomes, demonstrating their effectiveness in simulations and real-world prostate cancer data.

Contribution

It compares various machine learning approaches for variable selection in high-dimensional social-environmental data, identifying the most effective methods for true association detection.

Findings

Elastic net identified many true positives

Lasso controlled false positives well

Sparse group lasso and Bayesian trees showed strong performance

Abstract

Objective: Social-environmental data obtained from the U.S. Census is an important resource for understanding health disparities, but rarely is the full dataset utilized for analysis. A barrier to incorporating the full data is a lack of solid recommendations for variable selection, with researchers often hand-selecting a few variables. Thus, we evaluated the ability of empirical machine learning approaches to identify social-environmental factors having a true association with a health outcome. Materials and Methods: We compared several popular machine learning methods, including penalized regressions (e.g. lasso, elastic net), and tree ensemble methods. Via simulation, we assessed the methods' ability to identify census variables truly associated with binary and continuous outcomes while minimizing false positive results (10 true associations, 1,000 total variables). We applied the most promising method to the full census data (p=14,663 variables) linked to prostate cancer registry data (n=76,186 cases) to identify social-environmental factors associated with advanced prostate cancer. Results: In simulations, we found that elastic net identified many true-positive variables, while lasso provided good control of false positives. Using a combined measure of accuracy, hierarchical clustering based on Spearman's correlation with sparse group lasso regression performed the best overall. Bayesian Adaptive Regression Trees outperformed other tree ensemble methods, but not the sparse group lasso. In the full dataset, the sparse group lasso successfully identified a subset of variables, three of which replicated earlier findings. Discussion: This analysis demonstrated the potential of empirical machine learning approaches to identify a small subset of census variables having a true association with the outcome, and that replicate across empiric methods.