26–30 Jun 2022
Europe/Berlin timezone

Functional analysis of variance in presence of outliers: the RoFANOVA approach

27 Jun 2022, 15:00



Fabio Centofanti (University of Naples)


New data acquisition technologies facilitate the acquisition of data that may be described as functional data. The detection of significant changes in group functional means determined by shifting experimental settings, which is known as functional analysis of variance (FANOVA), is of great interest in a lot of applications. When working with real data, it's typical to find outliers in the sample, which might significantly bias the results. We present the novel robust nonparametric functional ANOVA approach (RoFANOVA) proposed by Centofanti et al. (2021) that decreases the weights of outlying functional data on the analysis outcomes. It is implemented using a permutation test based on a test statistic calculated using a functional extension of the traditional robust M-estimator. The RoFANOVA method is compared to several alternatives already present in the literature, using a large Monte Carlo simulation analysis, in both one-way and two-way designs. The RoFANOVA's performance is proven in the context of a stimulating real-world case study in additive manufacturing that involves the analysis of spatter ejections. The R package rofanova, which is available on CRAN, implements the RoFANOVA technique.

Centofanti, F., Colosimo, B. M., Grasso, M. L., Menafoglio, A., Palumbo, B., & Vantini, S. (2021). Robust Functional ANOVA with Application to Additive Manufacturing. arXiv preprint arXiv:2112.10643.

Keywords Functional analysis of variance; Functional data analysis; Functional M-estimators; Additive manufacturing

Primary authors

Fabio Centofanti (University of Naples) Prof. Bianca Maria Colosimo (Department of Mechanical Engineering, Politecnico di Milano, Milan, Italy) Prof. Marco Luigi Grasso (Department of Mechanical Engineering, Politecnico di Milano, Milan, Italy) Alessandra Menafoglio (Politecnico di Milano - Department of Mathematics) Biagio Palumbo (Università di Napoli Federico II) Simone Vantini (MOX - Dept of Mathematics, Politecnico di Milano, Italy,)

Presentation materials

There are no materials yet.