Use of comparative classification techniques to build a system for diagnosing heart diseases

Abstract

Heart diseases are among the most prevalent diseases, especially in our Yemeni situation, because of the lack of health culture in the community, lack of attention to health care, and costs of medical diagnosis in medical examination and consultation. Accurate and fast computer systems help to diagnose diseases, including heart diseases. In this article, heart disease symptom data were studied, analyzed, and classified using different classification techniques (Naive Bayes id3, j48, Decision Table, Cart, and ANN) to classify and diagnose common heart diseases and compare the results of the different analyses to see which provides better classification and diagnosis. The amount of distortion in the data used (training group) was calculated using entropy to build the decision tree. Many of the medical causes used in the analysis were reviewed through patient records and the symptoms of heart diseases helped in building and designing a web-based system that helps diagnose heart diseases. Some of the techniques mentioned earlier are simulated through the system interfaces, through which the patient can enter their pathological symptom data to make diagnosis.