<![CDATA[Current customer interest prediction is a very influential factor in the development of a business. Companies that are less effective in determining the quantity of merchandise inventory result in companies often making excessive purchases of goods. Excessive purchases of goods result in accumulation of goods so that financial flows do not run well and smoothly. Based on these problems, a data mining algorithm and calculation method are needed to predict future inventory and can help PT Macro Jaya Agung predict customer interest so that inventory remains stable. The classification algorithm used in data mining is K-Nearest Neighbor & C4.5, while the method used for calculations is safety stock & ROP (Reorder point). In the K-Nearest Neighbor & C4.5 algorithm using the RapidMiner application, the accuracy results for K-Nearest Neighbor are 88.89% using the confusion matrix and 86.00% using cross validation, while C4.5 has an accuracy rate of 66.67% using the confusion matrix and 71.81% using cross validation. The safety stock & ROP (Reorder point) method is used to determine the return order point for medical devices in order to determine the amount of stock of goods so that there is no buildup of goods and complaints from customers, so that the company's flow runs well and smoothly. The results of the 2.2L iron tube (0.25 M3) study were obtained for 131,889 pcs of safety stock and 186 pcs of ROP.]]>
