An automated medical diagnosis system based on symptom clustering approach

Abstract

Medical diagnosis is an important task that should be performed as accurately and efficiently as possible since a proper diagnosis may lead to increase the quality of life of the people. Educating patients about the symptoms will improve the chances of obtaining the right diagnosis and obtaining prompt and correct treatment. In many automated diagnosis systems, the results are influenced always by the bias of the researchers' initial assumptions. What is needed instead is an approach that minimizes human bias and considers all relevant data in determining a diagnosis. This paper reveals the design and implementation of IMedicare, a system that discover the most suitable and accurate diagnosis by extracting information on diseases. A proper clustering mechanism is being used to categorize diseases into groups according to the symptoms revealed. This system is proposed to accommodate several features such as predicting the most probable diseases through the symptoms, providing better communication among the doctors and the patients concealing the privacy and etc. Key areas of diagnosis are identified during the development process of the system in order to provide a better experience to the users through this approach. Analysis were done after consulting several medical doctors and based on their feedbacks. The proposed system was implemented using a dataset extracted from several doctors and the details from the web. Then the data set was segmented using a clustering approach and the system was evaluated using extensive datasets.