FORECAST SEASONING PRODUCT AT PT.SA USING TIME SERIES METHOD AND NEURAL NETWORK

Abstract

Over time, business competition among Industrial organizations develops rapidly. Therefore, it requires every company to always have the ability to compete among other companies. Most of the companies are trying to deal with the customers that have different degrees of demand variability. Companies often face fluctuations in customer orders, especially for seasoning products, to deal with these fluctuations the company needs to hold extra inventory, reserve production capacity, schedule overtime work plans and manage delivery scheduling times. It was become the role of a sales forecasting method. There are many method in sales forecasting, but the company does not know which methods can give the best results. In this study, compared between the time series and Neural Network method. Through three kind of data that is used, show that the best method which is Neural Network. It can be said as the best method because the result of error measurement that generated by Neural Network is smallest than others. Based on calculation Mean Absolute Deviation (MAD), Mean Squared Error (MSE) and Mean Absolute Percent Error (MAPE). For Ginger Powder the result error of MAD is 87.17, the result of MSE is 1078.30 and the result of MAPE is 16,56%. For Cassia Powder the result of MAD is 23.17, the result of MSE is 2375.67 and 11.504% for MAPE. The result forecast error of Black pepper Powder is 25.17 for MAD, the result of MSE is 1510.65 and 17.14% for MAPE. The result of Validation, Verification and Residual Test (IIDN) for the forecast result of ginger powder, cassia powder and black pepper powder using neural network is pass the validation test whether it is verification test or tracking signal test, normality test and auto-correlation test.