Variable Importance Based Interaction Modeling with an Application on Initial Spread of COVID-19 in China

TL;DR

This paper introduces a variable importance based interaction modeling (VIBIM) method for linear regression with both continuous and categorical predictors, improving stability and interpretability in high-dimensional settings, demonstrated through COVID-19 spread analysis.

Contribution

The paper proposes a novel VIBIM procedure that enhances interaction modeling stability and interpretability, especially for high-dimensional data with mixed predictor types.

Findings

VIBIM produces stable, interpretable models in simulations.

Application to COVID-19 data reveals relevant factors affecting spread.

VIBIM outperforms existing methods in stability and prediction accuracy.

Abstract

Interaction selection for linear regression models with both continuous and categorical predictors is useful in many fields of modern science, yet very challenging when the number of predictors is relatively large. Existing interaction selection methods focus on finding one optimal model. While attractive properties such as consistency and oracle property have been well established for such methods, they actually may perform poorly in terms of stability for high-dimensional data, and they do not typically deal with categorical predictors. In this paper, we introduce a variable importance based interaction modeling (VIBIM) procedure for learning interactions in a linear regression model with both continuous and categorical predictors. It delivers multiple strong candidate models with high stability and interpretability. Simulation studies demonstrate its good finite sample performance. We apply the VIBIM procedure to a Corona Virus Disease 2019 (COVID-19) data used in Tian et al. (2020) and measure the effects of relevant factors, including transmission control measures on the spread of COVID-19. We show that the VIBIM approach leads to better models in terms of interpretability, stability, reliability and prediction.