Design and Application of K-Means Method to Predict Sales at Arya Elektrik Stores

Abstract

In a shop, the product is a staple that is sold and bought. There are products in the store between products that sell well and products that don't sell. Given this problem, it is necessary to create a system that can classify products that sell, products that sell well, and products that don't sell well, which was carried out at the Arya Elektrik Store and carried out from March to July 2022. The K-Means algorithm is not affected by the order of objects used. used, this is proven when the author tries to randomly determine the starting point of the cluster center of one of the objects at the start of the calculation. The number of cluster memberships generated is the same when using another object as the starting point for the cluster center. However, this only affects the number of iterations performed. The purpose is to create applications and analyze product sales at the Arya Elektrik Store using the K-Means method. With this system, it can provide convenience benefits for analyzing the grouping of product sales at the Arya Elektrik Store, determining and classifying product sales that are selling well, very selling, and less selling. The method used to collect data is observation and interviews. With this application, shop owners can see the results of grouping these products. So, if there are products that don't sell well, shop owners can look for other alternatives so that products that don't sell can be sold.