Research on machine learning methods for recognition and classification of cardiovascular pathologies

TL;DR: A system that minimizes both medical and hardware errors in the interpretation of echocardiography and electrocardiography results using neural networks and machine learning methods is developed.

Abstract: This research describes machine learning methods and algorithms for early diagnosis of cardiovascular diseases. In a comparative analysis of the noninvasive diagnostic methods used, echocardiography and electrocardiography are the most common. This article explores new methods of automatic augmentation and correction of echocardiogram and electrocardiogram results using machine learning methods and algorithms. The article develops a system that minimizes both medical and hardware errors in the interpretation of echocardiography and electrocardiography using neural networks and machine learning methods. The scientific novelty of the study is that machine learning methods can reduce image analysis time, accelerate clinical decision making, and provide feedback to less experienced clinicians. The experiment, by training models of pathology recognition and classification, gave a clear idea of how to create the same model for other, more serious diseases, such as cancer and its various types. The purpose of this article is to analyze new methods for automatically adding and correcting echocardiogram and electrocardiogram results. The practical relevance of this study is to use computational resources to improve subsequent interpretation using machine learning algorithms.