This paper discusses a modification of the kappa measure of surprise and uses it to build semi-qualitative probabilistic networks. The new measure is designed to enable the definition of partial-order relations on its conditional values and is hence used to define qualitative influences over the edges of the network, similarly to Qualitative Probabilistic Networks. The resulting networks combine the advantages of kappa calculus of robustness and ease of assessment and the efficiency of Qualitative Probabilistic Networks. The measure also enables a built-in tradeoff resolution mechanism for the proposed network. © 2009 Springer Berlin Heidelberg.
CITATION STYLE
Ibrahim, Z. M., Tawfik, A. Y., & Ngom, A. (2009). Surprise-based qualitative probabilistic networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5590 LNAI, pp. 228–239). https://doi.org/10.1007/978-3-642-02906-6_21
Mendeley helps you to discover research relevant for your work.