References & Citations
Statistics > Methodology
Title: Directional Outlyingness for Multivariate Functional Data
(Submitted on 14 Dec 2016 (v1), revised 10 Jan 2017 (this version, v2), latest version 22 Apr 2018 (v5))
Abstract: The direction of outlyingness is crucial to describing the centrality of multivariate functional data. Motivated by this idea, we propose a new framework that combines classical depth with the direction of outlyingness. We generalize classical depth to directional outlyingness for both point-wise and functional data. We investigate theoretical properties of functional directional outlyingness and find that it naturally decomposes functional outlyingness into two parts: magnitude outlyingness and shape outlyingness which represent the centrality of a curve for magnitude and shape, respectively. Using this decomposition, we provide a visualization tool for the centrality of curves. Furthermore, we design an outlier detection procedure based on functional directional outlyingness. This criterion applies to both univariate and multivariate curves and simulation studies show that it outperforms all existing methods. Weather and electrocardiogram data demonstrate the practical application of our proposed framework.
Submission history
From: Wenlin Dai [view email][v1] Wed, 14 Dec 2016 12:55:51 GMT (516kb)
[v2] Tue, 10 Jan 2017 06:14:12 GMT (517kb)
[v3] Sun, 30 Apr 2017 10:53:17 GMT (335kb)
[v4] Sun, 17 Sep 2017 10:33:22 GMT (732kb,D)
[v5] Sun, 22 Apr 2018 06:48:35 GMT (1940kb,D)
Link back to: arXiv, form interface, contact.