Review of Methods for Functional One-Way Analysis of Variance

Accepted - September 2024

Abstract

There are many statistical methods for the functional analysis of variance. They are based on different concepts for example aggregating pointwise statistics, basis function expansion, graphical envelope, principal components analysis, and random projections. This paper concerns different aspects of such strategies. The paper covers, in particular, one-way and multi-way analyses, univariate and multivariate functional data, independent and repeated measurements, and fully and partially observed functional data. Theoretical constructions and properties are accompanied by summaries of simulation results and illustrated on real data examples. Additionally, the R code for performing the analyses in the examples is available on GitHub.