Predicting Relapse of the Myocardial Infarction in Hospitalized Patients

Abstract

The death pace of Myocardial Infarction is high throughout the world. The sickness happens because of the blockage of coronary arteries supply in the heart. Almost half of the individuals die before arriving at the clinic, and the recovery turns out to be problematic when the patient gets a second or third attack before admission of the hospital. Most indications displayed in the patients are cold sweats, windedness, and indigestion. If the diagnosis is not initiated in the following 5 - 6 hours, most of the cardiovascular muscles get destroyed, and the patient healing becomes severe. Indeed, even in joining the medical clinic, it is compulsory to give utmost priority to the patient for a week or at least for three days. The chances of healing become low when the patient gets repetitive attacks. The information gathered from the UCI ML repository will be used to predict the relapse of MI. Machine learning algorithms play a crucial part in predicting the relapse of MI during the first hours, second day and third day in a patient after admitting to the hospital. The dataset comprises information related to Age, sex, heredity on coronary heart disease, Increase of sodium in serum, serum sodium content, white platelet count, etc. The algorithms like AdaBoost (Adaptive Boosting), XGBoost (Extreme Gradient Boosting), Decision Tree, Random Forest, Support Vector Classifier (SVC) are used to anticipate the Relapse of the disease. By anticipating the reoccurrence of the illness ahead of time, medical specialists will be able to manage the time efficiently, and the medical practitioners can develop prescriptions to save the life of a patient without occurring serious conditions. SVC performed effectively in predicting the relapse of myocardial infarction on first and third day, while XGBoost performed effectively in predicting relapse of myocardial infarction on second day.

Topics & Concepts