affected by a small number of outliers, thus providing poor out-of-sample forecasts. This paper suggests a robust regression … approach,based on the S-estimation method, to construct forecasting models that are less sensitive to data contamination by …