Design of a New Hybrid Model using k-means Clustering and Two Class Neural Network for Heart Disease Prediction

Abstract

Heart diseases have become one of the most common and severe ailments to affect the well being of mankind in recent times. Approximately, 175 million deaths occur worldwide due to cardiovascular diseases. The healthcare industry generates millions and trillions of data which unfortunately is not efficiently reserved for mining and disease prediction purposes. The process of exploring relevant and intelligible information from extensive amounts of data is known as data mining. Clustering, Naive Bayes, decision trees, regression, artificial neural networks are some popular data mining techniques. It has been observed from the previous research that instead of applying a single mining algorithm, better results are obtained if a combination of mining algorithms can be applied. These combinational models, referred to as hybrid models, have set new trends in mining techniques. This study focuses on the predictive analysis of such a hybrid mining model that is a combination of k-means clustering and two class neural network on a heart disease dataset. The implementation of the project is done using Microsoft Azure Machine Learning Studio. Accuracy of the proposed model is 95.83%.