ANGULAR HALFSPACE DEPTH: CLASSIFICATION USING SPHERICAL BAGDISTANCES

Directional data lies on the surface of the unit sphere. Exploiting new results on the computation and the properties of the angular halfspace depth, we introduce the spherical version of the bagdistance, applicable to directional data. A bagdistance-based classification method for directional data is considered. The proposed method will be compared with other directional classifiers by means of a simulation study.