Extremality measures and a rank test for functional data

TL;DR

This paper introduces new extremality measures for functional data, providing a rank test that effectively captures growth pattern differences in children, applicable to complex high-dimensional data.

Contribution

It proposes novel extremality-based definitions and orderings for functional data, along with a practical rank test for analyzing complex observations.

Findings

The rank test detects different growth patterns between boys and girls.

Extremality measures are computationally feasible for high-dimensional data.

The methodology is applicable to microarray data and images.

Abstract

The statistical analysis of functional data is a growing need in many research areas. In particular, a robust methodology is important to study curves, which are the output of experiments in applied statistics. In this paper we study some new definitions which reflect the "extremality" of a curve with respect to a collection of functions, and provide natural orderings for sample curves. Their finite dimensional versions are computationally feasible and useful for studying high dimensional observations. Thus, these extreme measures are suitable for complex observations such as microarray data and images. We show the applicability of these measures designing a rank test for functional data. This functional rank test shows different growth patterns for boys and girls when it is applied to children growth data.