Heart Disease Prediction Using Logistic Regression

Abstract

Abstract Heart is the main organ of the human body. Heart disease is one of the disease that contribute most death in the world. It occurs when the heart is not functioning with full potential to pump the blood to all parts of the body. In this study, the researcher attempt to determine the significant variables between absence or presence of heart disease with others variables such as age, resting blood pressure, serum cholesterol, maximum heart rate achieved and others variables such as gender, chest pain type, resting electrocardiographic result, fasting blood sugar, oldpeak, slope, number of major vessels and thalassemia. The provided data covers 270 patients information. The three methods were applied to heart disease data which are were binary logistic regression (BLR) models, BLR models with least quartile difference (LQD) method and BLR models with median absolute deviation (MAD) method. After comparing among three methods, it is was found that the binary logistic with applied MAD model tend to be the best model with the highest percentage of accuracy. According to the final model, it is were shown that. Only chest pain type, number of major vessels and thalassemia is significant and positively associated to heart diseases. This kind of study is to gain public awareness about the most important factor that can lead to heart disease so that they can take prompt action or make a prevention to avoid this disease from occurring. This study also gather new information to the public to learn about logistic regression and robust method since it demonstrated step-by-step in this research. It will make easier for them to understand how this statistical method operates and gives benefit to all