The Use of Vicinal-Risk Minimization for Training Decision Trees

Created by W.Langdon from gp-bibliography.bib Revision:1.7615

@Article{Cao:2015:ASC,

• author = "Yilong Cao and Peter I. Rockett",
• title = "The Use of Vicinal-Risk Minimization for Training Decision Trees",
• journal = "Applied Soft Computing",
• year = "2015",
• ISSN = "1568-4946",
• DOI = "doi:10.1016/j.asoc.2015.02.043",
• URL = "http://www.sciencedirect.com/science/article/pii/S1568494615001507",
• abstract = "We propose the use of Vapnik's vicinal risk minimisation (VRM) for training decision trees to approximately maximise decision margins. We implement VRM by propagating uncertainties in the input attributes into the labelling decisions. In this way, we perform a global regularisation over the decision tree structure. During a training phase, a decision tree is constructed to minimise the total probability of classifying the labelled training examples, a process which approximately maximises the margins of the resulting classifier. We perform the necessary minimisation using an appropriate meta-heuristic (genetic programming) and present results over a range of synthetic and benchmark real datasets. We demonstrate the statistical superiority of VRM training over conventional empirical risk minimisation (ERM) and the well-known C4.5 algorithm, for a range of synthetic and real datasets. We also conclude that there is no statistical difference between trees trained by ERM and using C4.5. Training with VRM is shown to be more stable and repeatable than by ERM.",
• keywords = "genetic algorithms, genetic programming, Decision trees, Vicinal-risk minimisation, Decision trees, Classification",

}

Genetic Programming entries for Yilong Cao Peter I Rockett

Citations