Research on the heart attack prediction based on logistic regression

Abstract

Heart disease greatly threats human life and health. Through machine learning algorithms and models, computers can autonomously classify and predict data, thereby achieving analysis and prediction of unknown data. This research used Logistic Regression algorithm and 14 physical indicators from 302 patients to investigate heart attack. It found that because the correlation coefficient is greater than 0.4, the correlation that have to do with heart attack and what kind of chest pain it is, the maximum value of heart rate, whether have exercise induced angina and the ST depression was strong. The correlation have to do with heart attack and age, sex, number of main blood vessels, thalassemia is weaker. Correlation have to do with heart attack and cholestoral, the fasting blood sugar was the weakest. People who is older and men have more possibility to develop heart disease. The accuracy of the prediction is about 85.95%. The findings of this paper suggests that Logistic Regression algorithm will play a crucial role in preventing heart disease thus bringing better treatment outcomes to patients