Developing Knowledge Based System Using Data Mining Techniques For Diagnosis And Treatment Of Diabetes.

Abstract

Diabetes is a disease that affects the body’s ability to produce or use insulin. According to International diabetes Federation Ethiopia is one of the 32 countries in African region. 425 million people have diabetes in the world and more than 16 million people in the Africa Region; by 2045 it will be around 41 million. There were 2.567.900 cases of diabetes in Ethiopia in 2015.In the incident of Ethiopia different problems are observed in health care centers. From different perspectives, these problems are the scarcity of domain experts, practitioners, domain experts’ skills, health facilities etc. The general objective of this study is to design and develop prototype knowledge based system using data mining techniques for diagnosis and treatment of diabetes. In this study, to develop prototype knowledge base system using data mining techniques for diagnosis and treatment of diabetes is proposed by applying experimental research design. The researcher used domain expert knowledge as supplement of data mining techniques knowledge. To identify the best performance model for extracting the hidden knowledge (i.e. rules), three experiments for three classification algorithms were conducted these are J48, PART and JRip. Finally, the researcher decided to use the results of J48 classification algorithm in the development of the prototype knowledge based system using data mining for diagnosis and treatment of diabetes because it registered better performance than other classifiers. The developed model was tested with test performance of instance classified correctly and accuracy and only which performance score more than 95.1515 % accuracy was used as a knowledge base for the KBS development for a better efficiency and effective. So, Data Mining solves the knowledge acquisition problem of knowledge based perceptive by supplying extracted knowledge to KBS. Weka used for model construction and evaluation, Ultimate Visual basic studio 2013(Vb.net) for using data mining results as store knowledge base and as front side of prototype and common lisp prolog (Clisp) used for obtained knowledge backend coding. So, prototype knowledge based using data mining for diagnosis and treatment of diabetes was developed by integrating Vb.net and Clisp (i.e. clisp.net) tools. Finally, testing of the developed prototype KBS is done to evaluate the work of the system. The first one is testing in terms of using cases collected from hospital with that the system measured an accuracy of 92%. The second one is the user acceptance of the system is evaluated by the potential users’ of the system and achieved 91.43% recital. But further exploration has to be done to refine the knowledge base, languages, feature add etc. are need future advances.