Heart disease prediction model using tree-based method

Abstract

Heart disease has become a global trend illness, with high mortality and death toll. Based on the accumulated case data, data analysis can be used to predict potential heart disease patients more accurately and optimize the treatment. Comparing with other research, this study uses the most basic algorithm in the tree-based methods to predict the data. In the research, some exploratory analysis on the heart disease data set is done firstly. By analyzing the relationship between each variable, the symptom factors related to heart disease are obtained preliminarily. Then this study develops two heart disease prediction models using two kinds of tree methods and compares the error of the two models. It turns out that the random forest model fits better than the decision tree model which can predict the data much more accurate. This can help medical professionals predict heart disease status based on a patient's clinical data and improve patient treatment efficiency. Also, the study gives hints about the causes of the heart diseases, which will be useful in preventing these fatal illnesses.