Locally adaptative density estimation on Riemannian manifolds

GUILLERMO HENRY; ANDRES MUÑOZ; DANIELA RODRIGUEZ

SORT

INST ESTADISTICA CATALUNYA-IDESCAT

Lugar: Barcelona; Año: 2013 vol. 37 p. 111 - 130

1696-2281

In this paper, we consider a kernel type estimator with variable bandwidth when therandom variables belong in a Riemannian manifolds. We study asymptotic propertiessuch as the consistency and the asymptotic distribution. A simulation study is alsoconsider to evaluate the performance of the proposal. Finally, to illustrate the potentialapplications of the proposed estimator, we analyzed two real example where two dierentmanifolds are considered.