Neural Network Based Heart Disease Prediction

Heart attack is deadly disease which may result in sudden heart block, and it may lead to death. Every day many people are affected by a sudden myocardial infarction, and this may lead to sudden death. A heart attack is a medical emergency usually occurs when a blood coagulates or blood flow to the heart gets blocked and it may also occur due to poor contraction and expansion of cardiac muscles. Thus, we can able to understand that there must be a proper prediction method to predict the heart attack. The proposed system has been designed to predict the heart attack using Multi Layer Perceptron neural network-based classification approach. The objective of this project is to train a neural network and classify the type of heart attack into 17 categories such [Normal Sinus Rhythm, Auricular Premature beat, Atrial Flutter, Supraventricular Tachycardia, Atrial Fibrillation, Wolff-Parkinson-white syndrome, Premature ventricular contraction, Ventricular Bigeminy, Ventricular Trigeminy, Ventricular Tachycardia, Presbyopia, Succinate Dehydrogenase iron-sulfur subunit, Right bundle branch block, Fusion, Left bundle branch block]. Initially we have obtained an online data set [Physio net], by preprocessing and segmentation 4 different features are extracted. Then the extracted features are given as an input to neural network and classified into 17 classes.