Application of the K-Nearest Neighbor Algorithm in the Data Mining Process to Predict Drug Sales at Pratama Haji Medan-Pancing Clinic

Pratama Haji Medan-Pancing Clinic is a healthcare facility that routinely sells medications to patients. However, the current manual drug inventory management process poses risks such as delayed procurement and overstocking. To address this issue, this study aims to implement a data mining approach using the K-Nearest Neighbor (KNN) algorithm to predict drug sales at Klinik Pratama Haji Medan-Pancing. A quantitative research method was employed, utilizing historical drug sales data from the past two to three years. The data underwent a thorough process of assessment, cleaning, and transformation before being processed using the K-Neighbor Classifier from the scikit-learn library. The results demonstrated that the KNN method achieved a prediction accuracy rate of 88.9%, indicating its effectiveness in forecasting drug sales. By implementing this predictive system, Klinik Pratama Haji Medan-Pancing can improve the efficiency of inventory management, reduce the risk of stock shortages or surpluses, and support faster, data-driven decision-making. In conclusion, the KNN algorithm proves to be a feasible predictive solution for drug sales systems in clinics and holds potential for further development in intelligent and integrated pharmacy management.