Abstract:
The study was focused on an integration of expert system, mobile intelligence, and cloud for the diagnosis of diabetes type-1, type-2, and type-3. Diabetes occurred when a range of blood sugar is above normal (Fast Plasma Glucose >=100 mg/dL). Because the pancreas does not produce enough insulin or do not use insulin properly or both. Expert System is a program that uses knowledge base and inference engine to solve the problem in much more efficiently that are to require significant human expertise. Cloud has offered different services for developers to develop, manage, and deploy without fee and with fee to reduce the limitation of mobile- application. The number of people living with diabetes is increasing due to population growth, aging, addiction, urbanization, overweight, lack of physical exercise, and other complicated diseases. Moreover, these problems become worse due to the scarcity of specialists, miss diagnosis, and health facilities. Therefore, the diabetics need consistent treatment like dietary control, physical exercise and insulin management. To develop the prototype, the knowledge was extracted with semi-structure interview, and literature review, which are selected using purposive sampling technique from Aksum University and Dangla referral hospital because of accessibility of them. Then, the acquired knowledge is modeled using decision tree, and then represents using rulebased knowledge representation techniques and reasoning through hybrid chaining inference mechanism, which concludes a type of diabetes by checking the symptoms of diabetes through android studio and firebase platforms.
In testing and evaluating the prototype system, thirties diabetics’ cases are selected to test the accuracy of the prototype system and also ensure whether system satisfies the requirements of the end-users or not. Thus, the overall total performance of the prototype system is 81.65%. The reason why the prototype doesn’t record higher performance was an integrated system encountered some challenges during user acceptance testing like portability, usability, reliability, and functionality of the system. However, to make the system applicable in the domain area additional study is needed like diagnosis a disease with the user input instead of only the knowledge base, designed for other complicated diseases, large databases in cloud and data mining techniques.