Application of the K-Means Algorithm to determine the accuracy of goods inventory based on sales data classification at PT Batu Indah Raya

Abstract

The research aims to apply the K-means algorithm to classify sales data and to test the final results of Clustering Accuracy of each type of stone using DBI (Davies-Bouldin Index) testing. In the last 9 months PT Batu Indah Raya has experienced uncertain sales ups and downs, from these problems it is difficult to know the level of inventory of goods each month. The impact is that the company experiences a buildup or shortage of stock of goods every month and incurs costs related to excessive storage of goods or additional costs that arise due to lack of stock, to be able to overcome the problems that occur, requires a method or system to find out inventory by classifying sales data of each type and testing the level of accuracy with a test method or system. One method that can be used to solve these problems is using the K-means Clustering algorithm. The K-means method is expected to help companies identify inventory more easily. So that business people can show effectiveness in product / goods procurement. Data to determine the level of accuracy of inventory is taken based on sales every month at PT Batu Indah Raya. By using 2 clusters, namely cluster 1, it means that cluster 1 is in high demand and cluster 2 is less in demand in applying the k-means algorithm method with the results of the type of yellow brongkol stone in cluster 1 getting 59 members and cluster 2 getting 31 members with the DBI accuracy test result of 2.292 which is in the very good range. The type of white cobblestone in cluster 1 gets 36 members and cluster 2 gets 54 members with a DBI accuracy test result of 3.711 which is in the good range. The type of zeolite stone in cluster 1 gets 42 members and cluster 2 gets 36 members with a DBI accuracy test result of 3.597 which is in the good range.