Comparison of Forecasting Algorithms on Retail Data

Sales forecasting is of great importance in retail business in terms of reducing the number of stock days, stock cost and cash flow, increasing the availability of products in the stores, increasing sales and preventing customer loss. In this study, sales forecasting will be performed by using regre...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2022 10th International Symposium on Digital Forensics and Security (ISDFS) s. 1 - 4
Hlavní autoři: Dincoglu, Pelin, Aygun, Huseyin
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 06.06.2022
Témata:
Analytical models
Costs
data mining
Digital forensics
Error analysis
Predictive models
regression
Sales forecasting
time series
Time series analysis
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Sales forecasting is of great importance in retail business in terms of reducing the number of stock days, stock cost and cash flow, increasing the availability of products in the stores, increasing sales and preventing customer loss. In this study, sales forecasting will be performed by using regression and time series algorithms on the sales data of two stores of Migros, one of the largest retail stores in Turkey, for a period of two years. The two stores used in the experiments have different sales volumes. The monthly sales data received from Migros was first merged and then regression and time series algorithms were used to do forecasting. The error rates of the applied models were calculated for different scenarios and collected in a single table. By comparing the results of the two algorithms a decision was made for the choice of the better performing sales forecasting algorithm.
DOI:10.1109/ISDFS55398.2022.9800809