Modeling and forecasting product returns: An industry case study

The paper presents an industry case study on modeling and forecasting product returns in remanufacturing industry. Based on return patterns of 36 products over 20 months, Holt's and ARIMA methods are identified as the appropriate methods to forecast the future returns of 36 products. To decide which method is more accurate for forecasting the future return of each product, we use MAE, MAPE, and RMSE to evaluate the forecasting accuracy. The accurate method is recommended for forecasting each product.