Some Properties of the Gaussian Kernel for One Class Learning

13 years 11 months ago

Download www.cs.rpi.edu

This paper proposes a novel approach for directly tuning the gaussian kernel matrix for one class learning. The popular gaussian kernel includes a free parameter, σ, that requires tuning typically performed through validation. The value of this parameter impacts model performance signiﬁcantly. This paper explores an automated method for tuning this kernel based upon a hill climbing optimization of statistics obtained from the kernel matrix.

Paul F. Evangelista, Mark J. Embrechts, Boleslaw K

Real-time Traffic

Gaussian Kernel | Gaussian Kernel Matrix | ICANN 2007 | Popular Gaussian Kernel |

claim paper

» Complexity of Pattern Classes and Lipschitz Property

» Sparseness Versus Estimating Conditional Probabilities Some Asymptotic Results

» Learning Convex Combinations of Continuously Parameterized Basic Kernels

» Learning with Transformation Invariant Kernels

» Learning Translation Invariant Kernels for Classification

» On Representing and Generating Kernels by Fuzzy Equivalence Relations

» A DCprogramming algorithm for kernel selection

» Learning Gaussian Process Models from Uncertain Data

Post Info
More Details (n/a)

Added	08 Jun 2010
Updated	08 Jun 2010
Type	Conference
Year	2007
Where	ICANN
Authors	Paul F. Evangelista, Mark J. Embrechts, Boleslaw K. Szymanski

Comments (0)

Sciweavers

Some Properties of the Gaussian Kernel for One Class Learning

Gaussian Kernel | Gaussian Kernel Matrix | ICANN 2007 | Popular Gaussian Kernel |

Explore & Download

Productivity Tools

Document Tools

Image Tools

Sciweavers