Evaluation of the Optimal Timing of Diagnosis/Prognosis of Myocardial Infarction Using the MLP Artificial Neural Network

The prognosis of the risk of myocardial infarction is always an urgent and vital task in health care, in which ensuring the accuracy and timeliness of this diagnosis is especially important. For this purpose, the paper will present the use of Artificial Intelligence technology - MLP learning artificial neural network - to determine the optimal time in the diagnosis/prognosis of myocardial infarction. The myocardial infarction database used to train the network corresponding to time points on admission, at 24 h and 48 h after admission, were obtained from the UC Irvine Machine Learning Repository and preprocessed. Based on this database, 3 MLP neural networks corresponding to the 3-time points mentioned above were developed using the Deep Learning Toolbox of MATLAB software. The comparison of the results of training these networks showed that the time when predicting the risk of myocardial infarction after 24 h of hospital admission was the most accurate.