Representing Uncertainty in Medical Knowledge: An Interval Based Approach for Binary Fuzzy Relations

This paper addresses issues involved in representing uncertain causal relationships between entities in medical knowledge. An interval-based approach for medical binary fuzzy relations is proposed to represent the ignorance about uncertain and imprecise associative relationships whereas some effort has been put in introducing and reinterpretation of novel medical measures. The focus of attention of this model is based on utilizing interval-based fuzzy inclusion relationships as a causality measure expressing the strength of the degree of inclusion between fuzzy sets. Derived from the direction of an inclusion relationship an interval-based causal association can medically be interpreted as the necessity or the sufficiency of occurrence of a medical entity such as symptoms or disease with another one. Furthermore for simplification of computations and defuzzifications of dependent intervals a method for transformation of these relations into point-valued relations is proposed.