Evaluation of artificial intelligence based demand forecasting methods in terms of performance superiority: An application in the food sector

Businesses take various measures and decisions for the future uncertainties to protect and continue their existence under harsh competition conditions. These measures and decisions are shaped due to estimates made for both inside and outside the enterprise. In this regard, one of the significant forecasting types used in the measures and decisions that businesses will take against future uncertainty is demand forecasting. Considering the developments in demand forecasting methods used in the field of business, it is clear that methods based on artificial intelligence give more successful results than traditional methods. In this study, future demand forecasting models were generated by using the weekly sales data of frozen packaged kunafa of a firm in the Turkish food industry. In this forecasting methods based on artificial intelligence such as Artificial Neural Networks (ANN), Fuzzy Time Series and Adaptive Network Based Fuzzy Inference System (ANFIS) were employed. The performance comparison of the estimation models based on trial and error was made according to the Root Mean Square Error (RMSE) and Mean Absolute Percent Error (MAPE) criteria. As a result of the comparison, the demand forecast values for the future period were obtained based on the forecast model with superior performance