| Summary: | The incidence of myocardial infarction (MI) is growing year on year around the world. It is considered increasingly necessary to detect the risks early, respond through preventive medicines and, only in the most severe cases, control the disease with more effective therapies. The aim of the project was to develop a relatively simple artificial-intelligence tool to assess the likelihood of a heart infarction for preventive medicine purposes. We used binary classification to determine from a wide variety of patient characteristics the likelihood of heart disease and, from a computational point of view, determine what the minimum set of characteristics permits. Factors with the highest positive influence were: cp, restecg and slope, whilst factors with the highest negative influence were sex, exang, oldpeak, ca, and thal. The novelty of the described system lies in the development of the AI for predictive analysis of cardiovascular function, and its future use in a specific patient is the beginning of a new phase in this field of research with a great opportunity to improve pre-clinical care and diagnosis, and accuracy of prediction in preventive medicine.
|
|---|