Algorithms for bivariate zonoid depth (Q2456661)

Zonoid depth is a definition of data proposed by \textit{R. Dyckerhoff, G. Koshevoy} and \textit{K. Mosler} [Zonoid data depth: Theory and computation. COMPSTAT 1996. Proceedings in computational statistics. 12th biannual symposium, Barcelona, Spain, 1996. Heidelberg: Physica-Verlag. 235--240 (1996; Zbl 0894.62081)]. Efficient algorithms for solving several computational problems related to zonoid depth in 2-dimensional (bivariate) data sets are studied. The authors give the algorithmic results for the following problems related to zonoid depth: (1) Computing a contour, (2) Computing a depth map, (3) Testing the depth, (4) Computing the depth. Algorithms 2,3 and 4 are optimal. The results are obtained by exploiting an observation relating k-zonoids and k-sets.