IDENTIFICATION OF SALES SERIES WITH TREND AND SEASONALITY USING SELECTED METHODS
IDENTIFICATION OF SALES SERIES WITH TREND AND SEASONALITY USING SELECTED METHODS

Summary/Abstract: The actions taken by companies today are increasingly precise, dedicated to specific market expectations therefore require an increasingly solid basis for decision-making, especially at the strategic level. A key support in this area is forecasting methods that allow probing about the future with a certain probability. Therefore, methods for determining such forecasts are developing rapidly, cre- ating an arsenal of usable tools. However, many of them rely on modern information retrieval sys- tems, require accurate data with many variables and a long history. Despite the advancing digiti- zation, this is still unattainable for many (especially small) businesses. Therefore, the presentation and methods of short-term forecasting using analytical models should not be abandoned, because for many companies this is the best solution, and often the only one possible. This idea became the genesis of this article. It selects and presents methods dedicated to time series characterized by trend and seasonality, which are quite difficult to identify. They were compared and a method of model selection strategy was presented to select the forecast with the highest reliability. The models presented are easy to interpret and utilitarian and can provide effective support for supply chain management processes.

• Details
• Contents