Optimization of β-decision rules relative to number of misclassifications
MetadataShow full item record
AbstractIn the paper, we present an algorithm for optimization of approximate decision rules relative to the number of misclassifications. The considered algorithm is based on extensions of dynamic programming and constructs a directed acyclic graph Δ β (T). Based on this graph we can describe the whole set of so-called irredundant β-decision rules. We can optimize rules from this set according to the number of misclassifications. Results of experiments with decision tables from the UCI Machine Learning Repository are presented. © 2012 Springer-Verlag.
PublisherSpringer Science + Business Media
Conference/Event name4th International Conference on Computational Collective Intelligence, ICCCI 2012