Order Demand Forecast Using a Combined Approach of Stepwise Linear Regression Coefficients and Artificial Neural Network

Abstract
 Nowadays, businesses' forecasts to meet the demands have become more critical. This study aimed predict fifteen-day order demand for an fulfillment center using a Multilayer Perceptron Neural Network (MLPNN). The dataset used in was created from real database of large Brazilian logistics company and thirteen variables. Linear Regression Coefficients (LRC) were as feature selection method reduce estimation errors. showed that among variables, type_A (A5), type_B (A6), type_C (A7) had most significant impact on total forecasting. effect A6 found be greater than A7 A5. performance proposed model evaluated mean absolute percent error (MAPE). LRC-MLPNN provided MAPE 2.97%. results better forecasting obtained by selecting independent variables input with LRC. can also applied different problems.