A comparative study between the two prediction methods: holt-winter and the Arima model using census x13: a case study "al-Najah milk production company" in Maghnia

Abstract EN

The aim of this research paper is to compare the accuracy of the two prediction methods, Holt Winter and ARIMA using CENSUS X13 feature, in the case of a commercial production company, whose sales are affected by seasonality, which in our case is “Al-Najah Milk Production Company”, and with the help of the eviews 10 program, we got the forecast results for the company's sales for the year 2021 in both methods, and after comparing the prediction sales results with the real value of sales, we came to the conclusion that the ARIMA models using the census X13 property to correct seasonality and other influences, used in the American offices for studies, are the most accurate, where the sum of the squares of the differences between the real value of sales and the predicted sales results using ARIMA (census x13) it is smaller than its counterpart to the Holt Winter method, so the company can rely on the method in the future, for forecast it sales and even for make future decisions.