Abstract :
Congenital heart disease (CHD) is a disease with abnormalities in the structure of the heart or heart circulation function that has been brought from birth, CHD patients are increasing in number and occupy the largest position among heart disease in infants and children. In the incidence of CHD in Indonesia, it occupies a figure of 8 per 1000 live births, around 8090% of children with heart defects can survive to adulthood by undergoing treatment or surgery which has increased the patient's life expectancy. The dataset in this study is a combination of private data from patients with a history of congenital heart disease from Pusat Jantung Nasional Harapan Kita Hospital, and survey data google form with people who do not have a history of congenital heart disease. This research uses python jupyter notebook as software to predict congenital heart disease, with algorithm Artificial Neural Network Backpropagation as a method for predicting congenital heart disease, and MSE as a method of measuring the error value obtained. We tested 5 models patterns: 16-2-1, 16-3-1, 16-4-1, 16-5-1, and 16-6-1. The best results from the 5 models tested are the 16 input network model and 6 hidden layers and 1 output (16-6-1) with an MSE value of 0,04106, and an accuracy of 96%.
Keywords: Congenital Heart Disease, Prediction, Artificial Neural Network, Backpropagation.