In this paper, we consider decision trees that use both conventional queries based on one attribute each and queries based on hypotheses about values of all attributes. This approach is similar to one studied in exact learning, where membership and equivalence queries are considered. We propose dynamic programming algorithms for the minimization of the number of nodes in such decision trees and discuss results of computer experiments.
Bibliographical noteKAUST Repository Item: Exported on 2021-10-04
Acknowledgements: Research reported in this publication was supported by King Abdullah University of Science and Technology (KAUST). The authors are greatly indebted to anonymous reviewers for useful comments and suggestions.