Early Prediction of Diabetes Possibility using Boosting Decision Tree Pedi Nopedi (*), Aries Saifudin
Universitas Pamulang
Jl. Raya Puspitek No. 46 Buaran, Serpong, Tangerang Selatan, Banten, Indonesia, 15417
*pedi.nopedi1111[at]gmail.com
Abstract
Indonesia has a fairly high number of diabetic patients. Diabetes is a long-term chronic disease characterized by increased blood sugar (glucose) levels exceeding normal values. Diabetes that is not treated at an early stage can lead to complications of the disease. Diagnosing diabetes needs to be done to detect the disease early, so that the risk of diabetes can be anticipated early. Diabetes is influenced by many factors, so to make a diagnosis requires a complex analysis. An early prediction model for the possibility of developing diabetes using machine learning has been proposed. Some of the proposed models have not produced perfect prediction performance. Decision Tree is a simple classification algorithm but can provide a fairly good predictive model performance. Several publications show that Boosting technique can reduce single classifier algorithm errors. In this experiment, it is proposed to apply the Boosting technique to reduce the prediction model error based on Decision Tree. The experimental results show that the diabetes prediction model using the Decision Tree algorithm can be improved using the Boosting technique.
