OUTCOME PREDICTION FOR HEART DISEASE WITH ARTIFICIAL NEURAL NETWORK ALGORITHM

Abstract

Abstract — Heart disease is still the number one killer in the world, the latest discovery, this disease is the trigger of one third of all deaths in the world from year to year is always increasing. This study aims to help make predictions for heart disease early as well as create a model of prediction analysis (outcome prediction) obtained from health data (Healthcare). The method proposed in this study with deep learning techniques that apply artificial neural network algorithms with hidden layer techniques in making predictions of heart disease.  In this stage of research, problems were found in improving the accuracy of datasets used by handling problems in pre-processing data such as missing data and determining the form of data correlation. The model that was then tested through a heart disease dataset resulted in 90 % accuracy and with Random Forest result in 85% accuracy. With the creation of this prediction model is expected in addition to helping to make predictions of diseases can also provide the next innovation in data science in the field of health