An Expert System for Malaria Diagnosis using the Fuzzy Cognitive Map Engine

The complexity of malaria diagnosis increases because of symptom manifestation that could be confused with other tropical diseases and the fuzziness associated with patients' expression of their health conditions. There is a need for appropriate diagnostic tools that would assist the physician (or other trained medical personnel) in the differential diagnosis of malaria and other tropical diseases. In this paper, we present an initial result of an effort to develop a fuzzy cognitive map (FCM) system for the diagnosis of malaria. Concepts and their causality were defined based on the experiential knowledge from 30 physicians in 3 hospitals in Nigeria, who served as knowledge sources for this study. The semantic relationships among concepts were utilized in constructing an FCM model for malaria diagnosis, which was further integrated into a decision support engine (DSE). The comparative summary showed that the initial hypotheses (IH) by the physicians correctly matched the final diagnosis in 55% of the cases, whereas the accurate diagnosis (AD) of the FCM was 85%. This result is interesting; further analysis using Kendall's tau_b and the Spearman's rank order test also indicated a higher (though equally significant) correlation between the FCM results and AD than between IH and AD. The correlation between the physician's initial hypothesis and the FCM diagnosis was not significant.