Application of the Adaptive Boosting Method to Increase the Accuracy of Classification of Type Two Diabetes Mellitus Patients Using the Decision Tree Algorithm

Abstract

One of the data mining processes that is often used in machine learning is the data classification process. A decision tree is a classification algorithm that has the advantage of being easy to visualize because of its simple structure. However, the decision tree algorithm is quite susceptible to incorrect classification calculations due to the presence of noise in the data or imbalance in the data, which can reduce the overall level of accuracy. Therefore, the decision tree algorithm should be combined with other methods that can increase the accuracy of classification performance. Machine Learning is used through an artificial intelligence approach to solve problems or carry out optimization. Adaptive Boosting is used to optimize classification calculations. This study aims to examine the performance of Adaptive Boosting in the process of classifying second-degree diabetes mellitus patients using the Decision Tree algorithm. Diabetes mellitus is known as a chronic condition of the human body, the cause of which is an increase in the body's blood sugar levels because the body is unable to produce insulin or is unable to utilize insulin effectively, which is usually referred to as hyperglycemia.. By using a 60:40 data split, the Decision Tree algorithm produces an accuracy value of 95.71%, while the Adaptive Boosting-based Decision Tree results reach a value of 98.99%.