Design of a Diabetic Diagnosis System Using Rough Sets

Traditionally the diagnosis of a disease is done by medical experts with experience, clinical data of the patients and adequate knowledge in identifying the disease. Such diagnosis is found to be approximate and time-consuming since it purely depends on the availability and the experience of the medical experts dealing with imprecise and uncertain clinical data of the patients. Hence, to improve decision making with uncertain data and to reduce the time consumption in diagnosing a disease, several simulated diagnosis systems have been developed. Most of these diagnosis systems are designed to possess the clinical data and symptoms associated with a specific disease as knowledge base. The quality of the knowledge base has an impact not only on the consequences, but also on the diagnostic precision. Most of the existing systems have been developed as an expert system that contains all the diagnosis facts as rules. Notably, applying the concept of a fuzzy set has shown better knowledge representation to improve the decision making process. Therefore an attempt is made in this paper to design and develop such diagnosis system, using a rough set. The system developed is evaluated using a simple set of symptoms that is added to clinical data in determining diabetes and its severity.