PliableBVS: A flexible Bayesian variable selection method for modeling interactions with mandatory modifying variables
Researchers have introduced PliableBVS, a novel Bayesian variable selection method designed for complex interaction models. This approach extends the pliable lasso by incorporating spike-and-slab priors to induce sparsity, enabling simultaneous selection of main and interaction effects within a unified probabilistic framework. Simulation studies indicate that PliableBVS surpasses the original pliable lasso in identifying significant effects, reducing false discoveries, and enhancing prediction accuracy. Practical applications in studies related to labor onset and preeclampsia demonstrate its capability to pinpoint biologically relevant features and interactions. AI
IMPACT Introduces a more robust statistical tool for analyzing complex datasets, potentially improving feature selection in various scientific applications.