A New Approach of Dynamic Fuzzy Cognitive Knowledge Networks in Modelling Diagnosing Process of Meniscus Injury * *The present work was financially supported by the "Andreas Mentzelopoulos Scholarships University of Patras".

Abstract

Abstract A new approach of Dynamic Fuzzy Cognitive Knowledge Networks is presented. This is an evolutionary type of Fuzzy Cognitive Maps (FCM) that arose from the need for updating classic methodology in order to overcome its drawbacks, concerning the single calculation rule, stability and real time problems and expand its use in a variety of applications. This new approach is being tested for its accuracy in Decision Support Systems in medicine, trying to model knee injuries by using 17 real cases of patients. The new proposed model is able to diagnose meniscus injuries and to distinguish between acute and degenerative injury. Subsequently we observe the evolution of the injury by administering a proposed treatment by the physician. Results of this new method, which are presented in detail, are very satisfactory for both two levels and treatment stage, and in total agreement with Magnetic Resonance Imaging outcomes. The whole methodology is the outcome of a close collaboration between engineers and medical doctors and is significant because it is a promising tool which sets aside the main disadvantages of Fuzzy Cognitive Maps and allows us a wide use in many real time problems.