Sparepart sales clusterization and prediction using automatic clustering algorithm

Abstract

Abstract Creating an information system for a store becomes an important part in a business process as an effort to facilitate the running of the business process. Therefore, the existence of this information system will make it easier for the store to get information on products, recording the number of products entering and leaving, supplier information and can predict the number of sales in the coming month. Prediction of sales in the coming month is very important to facilitate the store in preparing stock of product in the coming month. The design of this program applies the automatic clustering method and fuzzy logical relationship that manages the historical data of the number of sales each month in the previous period, so we get a prediction of the number of sales in the coming month, which is useful for the store in preparing the amount of stock in the coming months. Based on the results of tests conducted, this program can provide predictions of the number of sales in the coming months. Tests are carried out on two product samples, where the products tested have an accuracy level of 86.20% and 90%, respectively.