In decision-making on optimal treatment strategies, it is of great importance to identify predictive variables that are involved in the decision rule, i.e. those interacting with the treatment. Effective variable selection helps to improve the prediction accuracy and enhance the interpretability of the decision rule. In this talk, I will present a new penalized regression framework which can simultaneously estimate the optimal treatment strategy and identify important variables. The advantages of the new approach include: (i) it does not require the estimation of the baseline mean function of the response, which greatly improves the robustness of the estimator; (ii) the convenient loss-based framework makes it easier to adopt shrinkage methods for variable selection, which greatly facilitates implementation and statistical inferences for the estimator. The new procedure can be easily implemented by existing state-of-art software packages like LARS. I will also present some theoretical properties of the new estimators and show their empirical performance using simulation studies and an application to an AIDS clinical trial.
Audience
This program is for the research community.