COMPSTAT 2016: Start Registration
View Submission - CRoNoS FDA 2016
A0167
Title: An angle-based functional pseudo-depth for shape outlier detection Authors:  Andre Rehage - TU Dortmund University (Germany) [presenting]
Sonja Kuhnt - Dortmund University of Applied Sciences and Arts (Germany)
Abstract: A measure especially designed for detecting shape outliers in functional data is presented. It is based on the tangential angles of the intersections of the centred data and its interpretation is the same as for a data depth. Due to its theoretical properties it is called functional tangential angle (FUNTA) pseudo-depth. Furthermore, a robustification called rFUNTA will be introduced. The existence of intersection angles is ensured through the centring. Assuming that shape outliers in functional data follow a different pattern than the regular data, the distributions of intersection angles differ from each other. Robustness properties of the two introduced measures and their performances in real data sets as well as simulation studies are investigated.