- title = "Evolution of Covariance Functions for Gaussian Process Regression using Genetic Programming",
- author = "Gabriel Kronberger and Michael Kommenda",
- howpublished = "arXiv",
- booktitle = "EuroCAST 2013",
- year = "2013",
- series = "LNCS",
- address = "Las Palmas, Canary Islands, Spain",
- month = "10-15 " # feb,
- publisher = "Springer",
- keywords = "genetic algorithms, genetic programming",
- bibdate = "2013-06-02",
- bibsource = "DBLP, http://dblp.uni-trier.de/db/journals/corr/corr1305.html#abs-1305-3794",
-
URL = "
http://arxiv.org/abs/1305.3794",
- size = "8 pages",
-
abstract = "In this contribution we describe an approach to evolve
composite covariance functions for Gaussian processes
using genetic programming. A critical aspect of
Gaussian processes and similar kernel-based models such
as SVM is, that the covariance function should be
adapted to the modelled data. Frequently, the squared
exponential covariance function is used as a default.
However, this can lead to a misspecified model, which
does not fit the data well.
In the proposed approach we use a grammar for the composition of covariance functions and genetic programming to search over the space of sentences that can be derived from the grammar.
We tested the proposed approach on synthetic data from two-dimensional test functions, and on the Mauna Loa carbon dioxide time series. The results show, that our approach is feasible, finding covariance functions that perform much better than a default covariance function. For the CO2 data set a composite covariance function is found, that matches the performance of a hand-tuned covariance function.",
-
notes = "EuroCAST 2013 not yet published (Aug 2013),
arXiv 22 May 2013 Also known as \cite{journals/corr/abs-1305-3794}
Presented at the Workshop Theory and Applications of Metaheuristic Algorithms, EUROCAST2013. To appear in selected papers of Computer Aided Systems Theory - EUROCAST 2013; Volumes Editors: Roberto Moreno-Diaz, Franz R. Pichler, Alexis Quesada-Arencibia; LNCS Springer",
