Relevance of User Data Completeness in Resolving Differential Diagnosis in Medical Expert System Optimization

Abstract

In this research we implemented a technique for resolving differential diagnosis (DD) in medical disease expert systems (ES). This was achieved by collecting additional information from users during diagnosis which help to correctly decide which new rule will be asserted to the working memory before feedback can be deduced after proving that a user is suffering from a particular disease by confirming all the symptoms. Our approach correctly resolves differential diagnosis by ruling out similar diseases which match user disease symptoms and enhances the general accuracy of the expert system due to sufficient user information made usable to the system. This work presents a general solution to most significant problem of differential diagnosis, which is neither restricted to a particular disease nor medical domain as compared to the work of other researchers where techniques to resolve differential diagnosis were restricted to a particular disease type. The system can be further optimized by providing the probability value when a goal is achieved as well as medical prescription after diagnosis is completed.