An Invariance Property of Predictors in Kernel-Induced Hypothesis Spaces

Authors:
Nicola Ancona;Sebastiano Stramaglia
Affiliations:
Institute of Intelligent Systems for Automation, C. N. R., Bari, Italy;TIRES, Center of Innovative Technologies for Signal Detection and Processing, University of Bari, Italy/ Dipartimento Interateneo di Fisica, University of Bari, Italy/ Istituto Nazionale di Fisica ...
Venue:
Neural Computation
Year:
2006

Citing 2
Cited 4

An equivalence between sparse approximation and support vector machines

Neural Computation
Kernel Methods for Pattern Analysis

Kernel Methods for Pattern Analysis

Semi-supervised learning by search of optimal target vector

Pattern Recognition Letters
Assessing Nonlinear Granger Causality from Multivariate Time Series

ECML PKDD '08 Proceedings of the European conference on Machine Learning and Knowledge Discovery in Databases - Part II
Association of genetic profiles to Crohn's disease by linear combinations of single nucleotide polymorphisms

Artificial Intelligence in Medicine
Causal relationships between frequency bands of extracellular signals in visual cortex revealed by an information theoretic analysis

Journal of Computational Neuroscience

Quantified Score

Hi-index	0.00

Visualization

Abstract

We consider kernel-based learning methods for regression and analyze what happens to the risk minimizer when new variables, statistically independent of input and target variables, are added to the set of input variables. This problem arises, for example, in the detection of causality relations between two time series. We find that the risk minimizer remains unchanged if we constrain the risk minimization to hypothesis spaces induced by suitable kernel functions. We show that not all kernel-induced hypothesis spaces enjoy this property. We present sufficient conditions ensuring that the risk minimizer does not change and show that they hold for inhomogeneous polynomial and gaussian radial basis function kernels. We also provide examples of kernel-induced hypothesis spaces whose risk minimizer changes if independent variables are added as input.