A Model For Heart Disease Prediction Based On Artificial Neural Network

The stored data in medical databases usually contains precious and hidden knowledge, which can be helpful in detection, prediction and treatment of sicknesses. Discovery and extraction of this knowledge are performed using data mining algorithms and resulted into the creation of intelligent systems titled medical decision support systems. This research offers a new prediction method for heart disease that combines three forward perceptron neural networks. In the proposed method at the first step, three perceptron neural networks are created separately and in the next step, the combined approach is done through voting. The experiment was performed on medical data of NEZAJA healthcare. Therefore, each of three individual neural networks along with other data mining algorithms such as decision trees and simple Bayesian have been executed separately on this data in the same condition, and the accuracy for heart disease prediction was calculated. The experimental results show the superiority of the proposed approach compared with other methods.