Comparison of heart disease classification with logistic regression algorithm and random forest algorithm
Firda Anindita Latifah, İsnandar Slamet, Sugiyanto Sugiyanto
Abstract
Heart disease is one of the deadliest diseases in the world. Many factors that trigger heart disease, such as age, sex, cholesterol, blood pressure, etc. Diagnosis of patients suffering from heart disease can be known from the symptoms experienced by the patient also the patient's and family's disease history. Further tests will be carried out to check for the actual disease. The length of the examination time will increase the risk borne by the patient. Therefore we need a classification algorithm that can speed up the process of diagnosing a disease. The Logistic Regression algorithm is one of the simple algorithms that have good classification capabilities as proven by previous studies. Logistic Regression develops the concept of Linear Regression with categorical dependent variables. Besides, there is also one algorithm that can be used in similar cases. Random Forest is an ensemble learning method, bagging from Decision Trees that works by randomly generating trees. This research compares the performance of the two algorithms in the classification of heart disease. This data research uses the Framingham Heart Study dataset. With define data into training and testing, the result of the classification shows that Logistic Regression is greater with an accuracy of 85.04% and 84.4% for Random Forest.