A Model of Analysis of Daily ECG Monitoring Data for Detection of Arrhythmias of the “Bigeminy” Type

The article presents the results of a study on the creation of a neural network model of a big data processing system obtained from daily ECG monitoring devices (Holter sensor) to search for characteristic points on the ECG graph. A method of overcoming the noise level of the initial data found in the course of the study is described. Using the example of the search for arrhythmias of the “bigeminia” type, the high efficiency of the model is shown, combining the use of a data preprocessing algorithm to eliminate noise and the use of a fully connected four-layer neural network. The presented results of a computational experiment to test the effectiveness of the model in organizing decision support by a cardiologist about the absence or presence of bigeminia are consistent with previously known theoretical provisions and experimental results in this area. As a result, the direction of further research for the creation of a decision support system by a cardiologist in the analysis of these technical means has been determined. #COMESYSO1120.