SEDPA, an expert system for disease diagnosis in eel rearing systems

The design, development and testing of a prototype interactive expert system (SEDPA) capable of diagnosing eel pathologies is described. The system incorporates a multiple subprogram modular design, although only the inference engine is largely described. Its user interface incorporates a natural language module. Starting from this kind of information, the system obtains its conclusions treating this information with a fuzzy controller and transmitting the uncertainty using the Dempster-Shafer theory (DST). The system's performance was evaluated in a series of tests. The results of a Fisher's exact test of the system's diagnoses versus those of the three fish pathologists for 29 eel pathologies indicated statistical differences in diagnostic performance with the two human experts. On the other hand, the use of different fuzzy associative memory (FAM) or representation of the different human experts's knowledge level indicated the system adaptability to the different work scenarios. The results described in this study have demonstrated the validity of this software for eel pathological diagnosis.