Application of Uncertain Temporal Relations Algebra to Diagnostic Problems

Temporal representation and reasoning has been applied so far to many areas of AI, including medical applications. Identification of temporal patterns plays an important role in medical diagnostics. Temporal scenarios of different diseases, therapy protocols, and other temporal graphs provide important additional information for medical decisions. In this paper we propose an application of algebra of uncertain temporal relations to diagnostic problems. We represent uncertain temporal relations within a medical scenario graph using the probabilities of the basic relations that can hold between two temporal primitives. Also in the paper we show how: (a) to generate temporal scenarios by integrating appropriate relational networks from already diagnosed cases; (b) to classify a new case using the measure of the distance between network and scenario. Key-Words: Temporal Diagnostics, Temporal Scenario, Temporal Relation, Uncertainty, Probability