부산대학교 도서관

Heart disease is quickly becoming one of the most critical and widespread health concerns in the globe at the present time. At this time, a great deal of research is being conducted in an effort to treat a variety of cardiac conditions and a great number of lives. Because it enables medical practitioners to begin treating patients as soon as possible, early diagnosis of any heart illness is critical for this purpose. The purpose of this research is to develop a method that is capable of detecting coronary disease even before the onset of symptoms. After acquiring a data set consisting of highly processed ECG signal levels and other parameters that are required for the diagnosis of a variety of heart-related conditions, anomalies and outliers were eliminated from the data. One part of the dataset consisted of inputs classified as a category, whereas the other part consisted of inputs classified as binary. After that, we used the Keras framework to construct a deep learning model with three layers, and then we fed it the data set. The accuracy of the model with the binary data set was 93.89 percent, which is satisfactory given the modification made to this data set to include cancer cases in adults aged 18 to 77. In addition to the artificial neural network, a K-Fold Cross Validation model was developed and data was input into it. The accuracy of this model in diagnosing heart illness was 91.6 percent, which is quite good and helps in recognizing anomalies. This model also had a high success rate in identifying heart disease. As a result of advancements in technology, we are now in a position to diagnose illnesses and abnormalities of the heart in a greater population with greater precision [ABSTRACT FROM AUTHOR]