.png){kind=icon}
.png){kind=icon}
Return
Reluctant Interaction Modeling in Generalized Linear Models
.png){kind=icon}
PRE
AI
G
DOI：10.1080/10618600.2026.2638491.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Including pairwise interactions in regression models can provide a more accurate approximation of theresponse surface. However, fitting such models remains particularly challenging in high-dimensional set-tings, where the number of interactions can reach millions or even billions. Although several methods havebeen proposed to address this issue, they typically rely on the hierarchical assumption or focus solely onlinear models with interactions. In practice, these assumptions are frequently violated. In this paper, weintroduce a flexible interaction modeling framework for generalized linear models that does not requirethe hierarchical assumption. Our approach extends the principle of interaction reluctance to generalizedlinear models, prioritizing main effects over interactions when their predictive contributions are similar. Themethod is easy to implement and scales well to very large datasets. We provide finite-sample guaranteesfor selection consistency in high-dimensional regimes. Numerical studies on simulated data and a realdataset demonstrate that our method achieves strong computational efficiency and favorable statisticalperformance.
Keywords:
Generalized linear models
High-dimensional interaction models
Variable screening
Journal
J
IF:
1.8
Papers:
116
Citations:
6.4K
Organization
Cited Papers