Cardiac disease related parameters analysis and disease prediction with health tips support using data mining and deep learning

  • Chetan Andhare, Musharraf Shaikh, Suyash Vartak, Saurabh Patil, Prasad Khawale, Dr. D. R. Ingale


Today the main reason behind deaths in the world is the heart disease. For citizens, medical experts and practitioners, the critical task is to predict the heart disease due to unawareness of symptoms that causes the major problem to the human being.  In India, awareness can be improved in the citizens and medical practitioners by providing deeper training and learning for the better human life. Despite heart function monitoring is the very essential extent of analysis and estimation that is the impact cause for the stage of heart with the other factors of heart healthiness like ECG, cholesterol, blood pressure and level of sugar. In previous research conducted on same issue, 13 features were used to predict heart disease. The results obtained in the previous work shows accuracy up to 93.33%. The research conducted by our group considering same features after changing the hyper parameters the accuracy was improved up to 95.25%. The basic objective of this is to develop a cost-effective application using deep learning technologies that will able to identify the cardiac disease severity for facilitating decision support system with health tips and suggestion. This paper explains the prediction system for monitoring heart fitness related parameters and described how to use deep learning algorithm and techniques like deep neural network to predict the heart disease in four stages.