PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Exploiting the Prior in the PAC-Bayes Bound
Emilio Parrado-Hernandez, John Shawe-Taylor and Amiran Ambroladze
(2007) Technical Report. Emilio Parrado-Hernandez, Madrid, Spain.

Abstract

This paper presents two SVM-like classification algorithms whose design criterion is to minimise the PAC-Bayes bound instead of to maximise the classification margin. A main goal of this work is to provide a good estimation of the generalisation capabilities of the algorithms, rather than just come up with new means to obtain a good (but unknown) classification rate in the test set. Some experiments illustrate the performance of these algorithms in comparison with the original SVM in a model selection plus classification task.

PDF - PASCAL Members only - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Monograph (Technical Report)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Learning/Statistics & Optimisation
Theory & Algorithms
ID Code:3267
Deposited By:Emilio Parrado-Hernandez
Deposited On:04 February 2008