Author(s)

Wei Ye, Arsen V. Melkumian

Corresponding Author

Wei Ye

ABSTRACT

Three models, including Naive Forecast, Seasonal Auto Regressive Integrated Moving Average (SARIMA), and artificial neural networks (ANNs), have been used to forecast the demand for the Red wine in Australia. The evaluation of the precision of each forecasting model is based on the Mean Absolute Percentage Error (MAPE). By comparison, it is found that ARIMA (1,1,5) (1,1,0)12 and ARIMA(1,1,6)(1,1,0)12 models to the red wine data and SARIMA models demonstrate the most superior performance.

KEYWORDS

Forecast, SARIMA, ANNs, MAPE