INVENTORY PREDICTION SYSTEM USING THE LEAST SQUARE METHOD AT THE SERBA SERBI ONLINE SHOP TANJUNGPINANG

Abstract

In this era of Industry 4.0, technology is advancing rapidly, and various applications are emerging to help traders optimize their product procurement. One such application is the forecasting or prediction system. Serba Serbi Online Shop Tanjungpinang is a shop that requires this system to manage its inventory successfully. As the shop offers a wide variety of goods, managing inventory has become challenging. The capital spent on inventory has increased, and unsold goods are piling up in the warehouse. To overcome this issue, the shop needs to calculate the amount of inventory required to reduce the stock build-up in the warehouse. The prediction method used for this purpose is the least square method. This method uses periodic or time-series data, which requires past sales data to predict future sales. Historical sales data is necessary to develop an inventory prediction system. The sales data for a single item from July 2022 to May 2023 was analyzed using the Least Square method. Based on this analysis, the predicted sales for June is 144.64, rounded to 145. The accuracy of the prediction results were tested, and the best accuracy was obtained from the prediction using 3 months of data, with an error rate of 5.47%. The prediction with the highest error rate was obtained from using 9 months of data, with an error rate of 10.65%. On average, the prediction accuracy is 0.02154%, which is considered very good.