IMPLEMENTASI NAIVE BAYES DALAM MEMPREDIKSI PENYAKIT DIABETES MELLITUS
Abstract
Diabetes mellitus is a illness that significantly affects the global population. This study explores the
implementation of Naive Bayes to predict diabetes mellitus. The problems faced include the complexity of
clinical datasets, feature diversity, and the need for accurate predictions. The proposed solution is to use
the Naive Bayes classification algorithm that utilizes a simple but strong assumption about feature
independence. The system is described with steps involving data pre-processing, dataset partitioning,
Naive Bayes model training, and performance evaluation. The datasets used include age, gender, weight,
HbA1C, fasting blood sugar. The results and testing show that the Naive Bayes model can provide fairly
accurate predictions for diabetes mellitus, with performance assessed through evaluation metrics such as ,
recall, accuracy, F1-score, and precision. In conclusion, the implementation of Naive Bayes is a fairly
effective approach to predicting diabetes mellitus. In this study, The performance of the Naïve Bayes
method is considered quite accurate, this is evidenced in the highest accuracy score of 85.71%. Despite its
simplicity, this algorithm can handle the complexity of the dataset and provide reliable predictions.
Keywords: Diabetes Mellitus, Data Mining, Machine Learning, Classification, and Naïve Bayes.