Розвідувальний аналіз даних для передбачення серцево-судинних захворювань методами машинного навчання

Abstract

The paper is devoted to the preparation and exploratory data analysis for further use in the information-analytical system of predicting cardiovascular disease risks and minimizing clinical risks using machine learning methods. A comprehensive analysis of the structure of the medical dataset "Heart Failure Prediction Dataset" was conducted, the distributions of categorical and numerical features were investigated, and the regularities of pathology development depending on age characteristics, chest pain type, exercise-induced angina, and specific ST segment changes on the electrocardiogram were revealed.