Functional Nonparametric Statistics in Action

Summary

Functional aspects are more and more frequent and varied in modern Statistics so much so that the designation of Functional Statistics had emerged recently. From a practical point of view, this is appearing as soon as one has to deal with data which are curves. A symbolical example of this new field of Statistics concerns the problem of nonparametric regression estimation in presence of functional data. This problem is doubly functional: the nature of the observed data (that is, the nature of the curves) is functional and the statistical model is also functional (that is, it is nonparametric). The only goal of this presentation is to show how recent Nonparametric Functional Regression Methods work in different practical situations. Several data sets are chosen to cover different fields of applied statistics (chemiometrics, speech recognition, econometrics) as well as different facets of statistical regression problems. Each example is quickly treated. Complete treatments, theoretical supports and extensive bibliographies are referred to other works.

The authors wish to thank all the participants of the working group STAPH on Statistique Fonctionnelle et Opératorielle of our department. Their continuous support and comments, through the activities of this group, are of great importance in the development of our researches. Complete activities of this group are available on line at http://www.lsp.ups-tlse.fr/Fp/Ferraty/staph.html.