Quantile Regression Based on Counting Process Approach under Semi-competing Risks Data

Abstract

Blocking is an important technique to reduce the noises introduced from uncontrollable variables. Many optimal blocking schemes have been proposed in literature but there is no single approach that can be applied to various blocked designs. In this article, we construct a mathematical framework using the count function (or the so-called indicator function) and develop a comprehensive methodology which allows us to select various optimal blocked orthogonal arrays: regular or non-regular designs with qualitative, quantitative or mixed-type factors of two, three, higher or mixed levels. Under this framework, most existing approaches are special cases of our methodology.?