Tropical Diseases Web-based Expert System Using Certainty Factor

Abstract

Indonesia is in the area of equatorial latitude which is known as a tropical climate country. On the other hand, many Indonesian people are prone to suffering typical tropical disease such as typhoid, dengue and malaria. With the motivation to curb disease, they should be informed about awareness, treatment and knowledge regarding to tropical diseases such as the symptoms, causes and early prevention. Web-based expert system is one well-known solution, which can access online via internet. Usually, traditional inference engine is possible to make misdiagnosis in medical domain. In addition, modern inference e.g. Bayes theory are complicated plus insufficient to substitute complete human brain reasoning activities. Therefore, this study aims to hybridize Forward Chaining and Certainty Factor method for diagnosing common tropical diseases suffered by Indonesian people. We have used ten types of tropical disease namely typhoid, dengue, chingkungunya fever, malaria, chicken pox, tuberculosis, diphtheria, pertussis, SARS and elephantiasis along with 38 symptoms for representing every disease into a knowledge base. We compare the results between expert’s diagnosis and our proposed web-based expert system after running ten times consecutively. We can conclude that hybridization of Forward Chaining and Certainty Factor methods during developing the web-based expert system, can significantly diagnose tropical diseases properly.