a recurrent rough polynomial artificial neural network and its biomedical application to the classification of a cardiac patient

Since the prevalence of cardiovascular disease and consequent mortality has increased, accurate prediction of the disease status of individuals is of great importance. therefore, models that have the least error and maximum reliability should be used. in recent years, the use of intelligent systems in engineering sciences and mechanics, especially in the prediction of diseases has increased dramatically. the diagnosis of disease by computer systems has become one of the important fields of study for researchers in this field. diagnosis of heart disease is an attractive and challenging field of research because of the high sensitivity of communities to the death and life of the patient. the use of medical information such as age, gender, blood pressure, blood glucose level, weight, blood cholesterol levels, bio-signal of electrocardiogram, etc. can help physicians in predicting heart disease to prevent the progression of the disease, recurrent heart attacks, and consequently, reduce mortality. this data should be collected in an organized manner and used to integrate the disease prevention and diagnosis system. evaluating these data and obtaining useful results and patterns in relation to it using data mining techniques and neural networks help to predict the early detection of this disease. this study presents a method of investigating factors influencing heart attacks using the recurrent rough group model of data handling (rrgmdh) neural network. we also compare the results of the proposed method to the results of five model of data handling neural network models, namely long short-term memory (lstm), gated recurrent unit (gru), redial basis function (rbf), probabilistic neural network (pnn) and recurrent group model of data handling (gmdh). the results indicate that the proposed method outperforms the five other methods.