Sales projection based on model of Holt-Winter

Abstract

Object: approbation of the Holt-Winters model or triple exponential smoothing, used for processes with a trend
and a seasonal component, to predict the supply of perishable goods on the example of flower products.
Methods: Holt-Winters model, ARIMA model, approaches to assessing economic efficiency.
Findings: models and methods for forecasting deliveries have been tested, and forecast values have been calculat-ed using the example of data on actual sales of the organization under study. A comparative analysis of the results of
modeling (forecasting) based on the Holt-Winters model and the ARIMA model showed that the prediction of time se-ries using the Holt-Winters model has reliable results by 99%, ARIMA has lower accuracy values (from 78 to 99.8%).
Conclusions: the methodology for making a qualitative forecast using the Holt-Winters and ARIMA models is
proposed for accurate planning of commodity flows and effective logistics coordination of the company's activities.
Based on the conducted research, a mathematical model was obtained for reducing the cost of storing excessive re-serves, recommendations for the use of forecasting models were proposed and the obtained forecast values were used to
assess the effect of the introduction of accurate forecasting methods. The use of the Holt -Winters model to predict the
supply of perishable goods on the example of floral products has a positive economic effect for the organization under study, equal to about 57% of net profit for the main items of floral products, which is achieved by reducing the cost of
transportation, storage, disposal of unrealized stocks.