A Dynamic Programming Algorithm for Learning Chain Event Graphs

Publication Type

Conference Proceeding Article

Publication Date

10-2013

Abstract

Chain event graphs are a model family particularly suited for asymmetric causal discrete domains. This paper describes a dynamic programming algorithm for exact learning of chain event graphs from multivariate data. While the exact algorithm is slow, it allows reasonably fast approximations and provides clues for implementing more scalable heuristic algorithms. © 2013 Springer-Verlag.

Keywords

chain event graphs, model selection, structure learning

Discipline

Theory and Algorithms

Publication

16th International Conference, Discovery Science 2013, Singapore, October 6-9, 2013. Proceedings

First Page

201

Last Page

216

ISBN

9783642408960

Identifier

10.1007/978-3-642-40897-7_14

Publisher

Springer Verlag

City or Country

Singapore

Share

COinS