Probability-Generating Function Kernels for Spherical Data

TL;DR

This paper introduces Probability-Generating Function kernels for spherical data, generalizing RBF kernels, and develops a semi-parametric learning algorithm to effectively analyze data on the unit hypersphere.

Contribution

The paper proposes PGF kernels supported on the hypersphere and a semi-parametric learning method, extending kernel techniques to spherical data analysis.

Findings

PGF kernels generalize RBF kernels for spherical data

Properties of PGF kernels are thoroughly studied

A semi-parametric learning algorithm is developed for these kernels

Abstract

Probability-generating function (PGF) kernels are introduced, which constitute a class of kernels supported on the unit hypersphere, for the purposes of spherical data analysis. PGF kernels generalize RBF kernels in the context of spherical data. The properties of PGF kernels are studied. A semi-parametric learning algorithm is introduced to enable the use of PGF kernels with spherical data.