Abstract

The influence of fast food is very large on people's health, it can cause dangers such as weight gain and obesity, poor nutrition, digestive disorders, increased blood pressure and cholesterol, increased risk of serious diseases such as cancer, diabetes and heart disease. Doctors who diagnose heart disease cannot be seen directly, with the help of an electrocardiogram, an examination using an electrocardiogram to detect abnormalities in the heart such as heart rhythm disturbances and pericarditis. The electrocardiogram tool detects how long the electrical waves last through measuring intervals and the amount of electrical activity in the heart muscle. The electrocardiogram tool consists of 12 leads. This research aims to diagnose heart abnormalities using the backpropagation method approach and to determine the level of accuracy of diagnosing heart abnormalities in patients. An artificial neural network is a network of interconnected nodes, where each connection has a connection weight distributed across all nodes. Meanwhile, the backpropagation learning algorithm is a guided learning algorithm whose input and output are known in advance. In addition to building artificial neural network models for predictive capabilities. Backpropagation is a method of working backwards, which can reduce the number of errors. The model will be evaluated using 10% separate data as validation data. The evaluation model created uses RMSE, confusion matrix, and ROV curve to find out how well the model performs in making predictions. The results of this research show an RMSE (Root Mean Square Error) value of 0.33, a matrix confusion value of 88.89%, and an ROC curve value with an ROC curve area = 0.88. The results of data analysis using artificial neural networks with the backpropagation method approach provide maximum results. The results of data analysis using artificial neural networks with a backpropagation method approach provide maximum results for diagnosing patient heart abnormalities.