A Novel MLP Technique on Augmented Dataset for Heart Attack Prediction

Abstract

Heart Attack is the leading life threatening disease because of the stress life facing by the humans. Traditional approaches are utilizing the 13 attributes related to medical tests like cholestral, threadmill test, and others. But the proposed model integrates the genetic algorithm known as “Particle Swarm Intelligence” algorithm with neural networks to perform the dimensionality reduction and classify the dataset using the customized neural network. The proposed model uses an augmented dataset that contains general information along with medical test reports, which are presented in the form of 55 attributes. The model also has the capability to predict the sub classification like mild, moderate, and severe. Neural Networks helps the automation system to extract the features automatically and implementation of genetic algorithm reduces the features and customization of layers helps the model to find the probability of each option and chooses the one with highest probability.