Detecting changes in the mean of functional observations

István Berkes, Robertas Gabrys, Lajos Horváth, Piotr Kokoszka

Journal of the Royal Statistical Society: Series B (Statistical Methodology) · 2009 · Volume 71 · Issue 5

DOI: 10.1111/j.1467-9868.2009.00713.x

Abstract

Summary.Principal component analysis has become a fundamental tool of functional data analysis. It represents the functional data asXi(t)=μ(t)+Σ1≤l<∞ηi,l+vl(t), whereμis the common mean,vlare the eigenfunctions of the covariance operator and theηi,lare the scores. Inferential procedures assume that the mean functionμ(t) is the same for all values ofi. If, in fact, the observations do not come from one population, but rather their mean changes at some point(s), the results of principal component analysis are confounded by the change(s). It is therefore important to develop a methodology to test the assumption of a common functional mean. We develop such a test using quantities which can be readily computed in the R package fda. The null distribution of the test statistic is asymptotically pivotal with a well-known asymptotic distribution. The asymptotic test has excellent finite sample performance. Its application is illustrated on temperature data from England.

Keywords

Change-point detection, Functional data analysis, Mean of functional data, Significance test

Enable JavaScript for citation tools and additional paper interactions.