Implementation of Product Sales Forecast Using Artificial Neural Network Method

Abstract

Product sales forecasting is used by companies to estimate or predict future sales levels using sales data in the previous year. The Artificial Neural Network Backpropagation Algorithm can forecast the sales of goods for the next period for each item in the company. The forecasting process begins by determining the variables needed in the network pattern, and then the established network pattern continued in the network training process using the backpropagation algorithm. After carrying out the network training process, the researcher comparisons with several network patterns formed. This research was conducted to discuss the forecasting analysis of PT XYZ products on spiral and leaf springs. Forecasting carried out on Toyota 48210-25290 R3 type leaf springs using the Artificial Neural Network Backpropagation method with a learning rate weight value of 0.1 hidden layers four and an error of 0.01. From the data processing analysis that has been carried out based on the weight parameters selected, the prediction of sales in April.