Carnegie Mellon University
Browse

Kernel Conjugate Gradient

Download (107.41 kB)
journal contribution
posted on 2005-01-01, 00:00 authored by Nathan Ratliff, J. Andrew Bagnell
We propose a novel variant of conjugate gradient based on the Reproducing Kernel Hilbert Space (RKHS) inner product. An analysis of the algorithm suggests it enjoys better performance properties than standard iterative methods when applied to learning kernel machines. Experimental results for both classification and regression bear out the theoretical implications. We further address the dominant cost of the algorithm by reducing the complexity of RKHS function evaluations and inner products through the use of space-partitioning tree data-structures.

History

Publisher Statement

All Rights Reserved

Date

2005-01-01

Usage metrics

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC