High-Dimensional Interaction Detection With False Sign Rate Control

Identifying interaction effects is fundamentally important in many scientific discoveries and contemporary applications, but it challenging since the number of pairwise interactions increases quadratically with covariates that higher-order grows even faster. Although there a growing literature on detection, little work has been done prediction false sign rate detection ultrahigh-dimensional regression models. This article fills such gap. More specifically, this we establish some theoretical results selection for quadratic models under random designs. We prove examined method enjoys same oracle inequalities as lasso estimator further admits an explicit bound rate. Moreover, can be asymptotically vanishing. These new characterizations are confirmed by simulation studies. The performance our proposed approach illustrated through real data application.