Publication

Abstract : Predictive analysis plays a major role in healthcare industry where forecasting the disease will reduce the risk that happen to patients. Statistics show that cardiovascular diseases have increased the mortality rate in India. Machine learning which is used in developing a predictive model for various domains is nowadays applied in the field of medical diagnostics. Machine learning is playing an integral role in predicting the presence or absence of heart diseases. Such predictions, if done well in advance, can help the doctors to carry out the treatment for the patients and mitigate their health risk. Biological samples such as blood or tissues are collected from the human body to predict cardiovascular diseases. The proposed work is focused on developing various machine learning predictive models using support vector machine, decision tree, neural network and K-nearest neighbour for prediction of heart disease. For this work Cleveland heart disease dataset is used which consists of 14 attributes and 294 records. A comparative analysis on the prediction models were carried out. From the results, it was found that support vector machine, decision tree and KNN (k = 15) classifiers yield better accuracy to predict heart disease than the other models.