Title: Forward screening for high dimensional additive quantile regression
Authors: Daoji Li - California State University Fullerton (United States) [presenting]
Abstract: A new feature screening approach is presented for high dimensional additive quantile regression. Under certain regularity conditions, we show that the proposed method all the important variables can be identified in a small number of steps. To remove noise variables after the screening step, we further implement variable selection via a modified Bayesian information criterion. We show that the smaller selected set still contains all the important variables with overwhelming probability. The method and theoretical results are supported by several simulations and real data examples.