The package: nonparametric regression using local rotation matrices in

Abstract

ABSTRACT The package implements nonparametric (smooth) regression for spherical data in , and is freely available from the Comprehensive Archive Network (CRAN), licensed under the MIT License. It can be used for regression when both the response and explanatory variables lie on the unit sphere. The model uses a flexible kernel-type regression determined by a rotation which depends on a smoothing parameter as well as the prediction point. A particular kernel is proposed and a smoothing parameter selection procedure is also provided. Finally, some examples are included in the package.