High-dimensional \(A\)-learning for optimal dynamic treatment regimes (Q1650064)

The main goal of this paper is a penalization \(A\)-learning method [\textit{S. A. Murphy}, J. R. Stat. Soc., Ser. B, Stat. Methodol. 65, No. 2, 331--366 (2003; Zbl 1065.62006)] by means of the Dantzig selector [\textit{E. Candès} and \textit{T. Tao}, Ann. Stat. 35, No. 6, 2313--2404 (2007; Zbl 1139.62019)]. The authors estimate the baseline mean function and the propensity score model with NP-dimensionality of covariates. The error bounds are derived for the corresponding estimates under the model misspecification. Some examples of application obtained results are given.