<![CDATA[Diabetes mellitus is a major global health issue with a steadily increasing prevalence in Indonesia, including the Surakarta region. Diabetes complications are among the leading causes of mortality and reduced quality of life in patients, while also imposing a significant economic burden. This situation highlights the need for an accurate predictive model to identify both high- and low-risk patients, particularly in primary healthcare facilities that still face limitations in conducting comprehensive risk prediction. This study aims to develop a predictive model for diabetes complications by optimizing the adaptive integration of the Naïve Bayes algorithm with Information Gain for feature selection. The dataset used includes demographic variables (age, gender), clinical data (blood glucose level, HbA1c, blood pressure), medical history (family history of diabetes), and lifestyle factors (physical activity, dietary patterns). The results indicate that the pure Naïve Bayes algorithm achieved an accuracy of 75%. After applying Information Gain for feature selection, the accuracy improved to 87.5%, representing a 12.5% increase. These findings demonstrate that integrating Naïve Bayes with Information Gain can produce a more accurate prediction model for diabetes complications, making it a potentially effective decision-support tool for healthcare professionals in preventing complications and reducing the burden on the healthcare system.]]>
Copyrights © 2025
