Classification of Diabetes Disease Using the Naive Bayes Algorithm

Abstrak

Diabetes is one of the leading causes of death worldwide, thus requiring an accurate early detection method to minimize the risk of complications. The urgency of this study is based on the need for a computational approach capable of identifying diabetes risk more quickly and objectively. The Naïve Bayes method was selected due to its simple yet effective classification capability in processing both numerical and categorical data. The dataset used was obtained from the Kaggle platform, consisting of 995 records, and was processed through several stages, including preprocessing, splitting the data into 80% training and 20% testing, modeling using the Naïve Bayes algorithm, and evaluation through RapidMiner and manual calculations. The results show that the model achieved an accuracy of 89.95% and an AUC value of 0.968, indicating excellent predictive ability in distinguishing diabetic and non-diabetic cases. Although this study is limited to only two attributes, the findings demonstrate that Naïve Bayes can serve as the basis for a fast and efficient early diabetes detection system. Overall, this research highlights the importance of applying data mining techniques to support the early diagnosis of diabetes