Application of Artificial Neural Network for Forecasting Demand Bottled Drinking Water by Using Back propagation Algorithm

Abstract

The erratic demand for bottled drinking water (BDW) products caused the sales target not to be achieved for several periods. One of the efforts that can be made by the management so that the amount of production is correct is by forecasting demand. This study aims to determine the best forecasting model using the Artificial Neural Network (ANN) method with the Backpropagation algorithm, supervised learning. The activation function used is the binary sigmoid function (logsig). Based on the result, the best architectural model is found in neurons 3-4-1 with an MSE value of 0.0002 and a MAPE value of 2.346%.