Application of the Naive Bayes Algorithm in Data Mining for Predicting Stroke Disease

Abstract

Stroke is a major global health problem that may lead to permanent disability or death. In Indonesia, analytical approaches such as data mining have been increasingly used to support early identification of stroke risk based on patient health records. This study applies the Naive Bayes algorithm as the main classification method. The research procedure includes data collection from the Kaggle repository, data selection based on predetermined clinical criteria, data cleaning to remove duplicates and missing values, and data transformation by converting categorical attributes into numerical form. The dataset was then split into training and testing subsets for model development. The final dataset consisted of 3,256 patient records containing variables such as gender, age group, hypertension, heart disease, average glucose level, body mass index (BMI), smoking status, and stroke occurrence. After completing these preprocessing stages, the Naive Bayes model achieved an accuracy of 89.49%. This result indicates that the model was able to classify stroke and non-stroke cases in the dataset with a satisfactory level of accuracy. The findings suggest that the model may serve as a baseline component for developing decision-support systems aimed at facilitating early identification of stroke risk in targeted population groups.