ANN, ARIMA and MA timeseries model for forecasting in cement manufacturing industry: Case study at lafarge cement Indonesia — Aceh

The accurate demand forecast method is one of the main important to industry to minimize error. In this study tried to propose the Artificial Neural Network (ANN), Arima and Moving Average (MA) to predict the condition of sale demand in cement manufacturing industry. The predicted months after the twenty two at the last months data and should be validated with the real two months data. The processes come from collecting sales real data from cement industry in aceh province. Analyzed the predicted condition and the mean square error (MSE), MAPE and SSE. Compared to the installed method in the factory should be also considered. The result of this study ANN, Arima and MA models are better than the installed method and the predicted data are better as well where the installment produce more than thirty percent errors.