| Abstrakt: |
Cardiovascular disease is a term used to describe a variety of heart diseases, illnesses and events that affect the heart and circulatory system. The aim of this study is to design a fuzzy expert system for heart disease diagnosis. The neuro-fuzzy system was designed with eight input fields and one output field. The input variables are heart rate, blood pressure, age, cholesterol, chest pain type, blood sugar, exercise and sex. The output detects the risk levels of patients, which are classified into four different categories: very low, low, high and very high. The dataset used was extracted from the UCI machine learning repository and preprocessed in order to make it appropriate for the training; then the initial FIS was generated. The network was trained with the set of training data, after which it was tested and validated with the set of testing data. The output of the system was designed in such a way that the patient could use it personally. The patient just needs to supply appropriate values which serve as input to the system, and based on the values supplied, the system will be able to predict the risk level of the patients. Fuzzy Interference System (FIS) is a combination of neural networks and adaptive neuro fuzzy. The results obtained from the system are compared with the SVM classifier. The system has been tested and the result showed over 90% accuracy. It has been shown that neuro-fuzzy is suitable and feasible to be used as a supportive tool for disease diagnosis. [ABSTRACT FROM AUTHOR] |
