<![CDATA[The rapid development of information technology encourages various sectors to adopt data-based approaches in decision-making, including the health sector. Apotek Tanikha faces challenges in identifying fast moving and slow moving drug categories accurately and systematically, as inventory management is still carried out manually based on the pharmacist’s experience. This study aims to apply the Naive Bayes algorithm to predict drug categories based on sales transaction data, measure the accuracy level of the algorithm, and produce operational inventory management recommendations. The research uses a descriptive qualitative approach by analyzing 2,949 sales transaction records from January to March 2026. Data preprocessing includes data cleaning, feature selection, aggregation, and labeling. The Naive Bayes algorithm is implemented using Google Colaboratory with an 80:20 training-testing data split (2,359 training records and 590 testing records). Model performance is evaluated using a confusion matrix, resulting in an accuracy of 98.11% precision of 98.26%, and recall of 98.11%. In conclusion, the Naive Bayes prediction model effectively categorizes drug inventory, providing concrete recommendations for Apotek Tanikha to optimize working capital, minimize drug expiration risks, and ensure the availability of fast moving drugs.]]>
Copyrights © 2026
