ANÁLISE COMPARATIVA DE PREVISÕES DE VENDAS: O MODELO HOLT-WINTERS X PREVISÃO REALIZADA

Abstract

In a business environment, with constant and increasingly rapid changes, the forecasting methods have helped in guidance on the future direction. In this sense, the present research deals with the application of a computational model, NCSS software (Number Cruncher Statistical System), to carry out the sales forecasts of a toy manufacturing company. The implementation of the Holt-Winters multiplicative model, surveys and preliminary analyzes of forecasts data from past sales of the five major products, prioritized by ABC classification, simulated data to compare forecasts with sales that occurred were performed, and designed the prediction of products prioritized for the next three years. After completion of the case study, there were good results of accuracy of the passed data of forecasts made, where the representativeness of the forecasts generated by the NCSS software exceeded the values from 0,75 to R², which validated the methodology.