Heart Disease Diagnosis Using Diverse Neural Network Categories

This study indicated an intelligent diagnosis method about the heart disease based on the diversity of artificial neural networks algorithms. The major objective was to improve the heart disease diagnosis accuracy, and reducing the miss-diagnosis results. One of the disadvantages of the current heart disease diagnosis methods that the diagnosis is expensive and not accurate enough to confirm the diagnosis of the heart disease. Our study will help the medical doctors to make appropriate diagnosis process and thus take specific treatment of the heart diseases easily. Our suggested method used Heart Diseases Database (Cleveland database) for training and testing phases. We used five neural networks: Quick, RBFN, Multiple, Prune, and Exhaustive. Our results revealed that the dynamic neural network algorithm was the best method diagnosing heart diseases accurately. We recommend using Dynamic neural network algorithm as the routine and sensitive method for diagnosis of heart disease easily and specifically.