Applying superquantile regression to a real-world problem: Submariners effort index analysis

Abstract

We lay out the fundamental theory for superquantile regression. Such novel regression framework is centered on a coherent and averse measure of risk, the superquantile (also called conditional value-at-risk), which yields more conservatively fitted curves than classical least squares and quantile regressions. We illustrate this regression technique by analyzing a real-world problem where a random variable represents the effort index of the Portuguese Navy submariners along their Navy careers. This index was created as a decision tool to support human resource management inside the Submarine Squadron.