|
AUTHOR(S): Uraz Corekci, Mehtap Dursun
|
TITLE Time Series Forecasting for Tobacco Product Sales Employing SARIMA, ETS, and TBATS Models |
 PDF |
ABSTRACT In this study, performance of three different time series are examined for predicting the monthly sales of tobacco products in the FMCG industry. Key forecast accuracy metrics, such as MAE, MSE, MASE, and sMAPE, are used to compare and assess each model using actual anonymized sales data. While ETS offers reliable performance with a more straightforward structure, the SARIMA model successfully captures regular seasonal patterns. Although it can occasionally result in overfitting, TBATS shows great versatility when modeling multiple or non-integer seasonalities. Python libraries like statsmodels and tbats are used to implement all models in the Google Colab environment. The results leads us in the selection of suitable forecasting techniques for stock management and operational planning in highly regulated and seasonal industries such as tobacco. |
KEYWORDS Time Series Forecasting, SARIMA, ETS, TBATS, Tobacco Sales, FMCG Sector, Forecast Accuracy Metrics |
|
Cite this paper Uraz Corekci, Mehtap Dursun. (2025) Time Series Forecasting for Tobacco Product Sales Employing SARIMA, ETS, and TBATS Models. International Journal of Economics and Management Systems, 10, 270-273 |
|
