The Consistency of the CADIAG-2 Knowledge Base: A Probabilistic Approach
The paper presents the methodology and the results of checking consistency of the knowledge base of CADIAG-2, a large-scale medical expert system. Such knowledge base consists of a large collection of rules representing knowledge about various medical entities (symptoms, signs, diseases...) and relationships between them. The major portion of the rules are uncertain, i.e., they specify to what degree a medical entity is confirmed by another medical entity or a combination of them. Given the size of the system and the uncertainty it has been challenging to validate its consistency. Recent attempts to partially formalise CADIAG-2’s knowledge base into decidable Gödel logics have shown that, on that formalisation, CADIAG-2 is inconsistent. In this paper we verify this result with an alternative, more expressive formalisation of CADIAG-2 as a set of probabilistic conditional statements and apply a state-of-the-art probabilistic logic solver to determine satisfiability of the knowledge base and to extract conflicting sets of rules. As CADIAG-2 is too large to be handled out of the box we describe an approach to split the knowledge base into fragments that can be tested independently and prove that such methodology is complete (i.e., is guaranteed to find all conflicts). With this approach we are able to determine that CADIAG-2 contains numerous sets of conflicting rules and compute all of them for a slightly relaxed version of the knowledge base.
KeywordsKnowledge Base Description Logic Column Generation Medical Entity Propositional Language
Unable to display preview. Download preview PDF.
- 1.Adlassnig, K., Kolarz, G., Scheithauer, W., Grabner, H.: Approach to a hospital-based application of a medical expert system. Informatics for Health and Social Care 11(3), 205–223 (1986)Google Scholar
- 5.Ciabattoni, A., Vetterlein, T.: On the fuzzy (logical) content of Cadiag2. Fuzzy Sets and Systems (2009) (to appear shortly)Google Scholar
- 6.Picado Muiño, D.: The (probabilistic) logical content of cadiag2. In: Proceedings of ICAART 2010, pp. 28–35 (2010)Google Scholar
- 11.Ciabattoni, A., Rusnok, P.: On the classical content of monadic G~ and its applications to a fuzzy medical expert system. In: Proceedings of the 12th International Conference on the Principles of Knowledge Representation and Reasoning (2010)Google Scholar
- 12.Klinov, P., Parsia, B.: Pronto: A practical probabilistic description logic reasoner. In: International Workshop on Uncertainty in Description Logics (2010)Google Scholar
- 20.Sattler, U., Schneider, T., Zakharyaschev, M.: Which kind of module should I extract? In: Grau, B.C., Horrocks, I., Motik, B., Sattler, U. (eds.) Description Logics. CEUR Workshop Proceedings, CEUR-WS.org, vol. 477 (2009)Google Scholar