Timely diagnosis of the heart disease using data mining algorithm

In the current state of medical knowledge, we are facing abundant data collection about various diseases. Investigating these data and obtaining useful results and patterns in conjunction with diseases, is the main propose for using them. In this paper, considering the significant share of the heart disease in human mortality, a variety of data mining methods and algorithms are applied for timely diagnosis of heart diseases. For this purpose, Dezful Hospital standardized data collection for patients with heart disease, have been used. Most important variables of this collection are exercise-induced angina, type of chest pain, age, maximum heart rate and blood pressure at rest. Classification algorithms of K-NN, PSO, Bagging and SVM + PSO hybrid algorithm, are tested and evaluated on the data sets. Based on our evaluations, the proposed hybrid algorithm, with accuracy of 94.74, had the highest accuracy.