Aplicação da metodologia Box & Jenkins na previsão das vendas de leite de um laticínio em Guaíra-PR

Abstract

Milk is considered of fundamental importance for the agricultural sector and in accordance with Chaddad (2015) milk production in Brazil needs to improve substantially in quality and quantity. Data from the United Nations Food and Agriculture Organization (FAO) indicate that between 2003-2013 the Brazilian milk production increased 49.3%, from 22.943 billion liters, the sixth largest in the world, to 34,255,000, this increase was mainly due to the expansion of cattle and to a lesser extent productivity increase. The data indicate that production should grow at an annual rate of 1.9%, and consumption is expected to increase at an almost equal rate of production. This growth rate of output is higher than that observed for the growth of the population. This study examined the performance expected from the sale of milk in a dairy located in Guaira – West Paranaense in the periods from 2004 to 2010, analyzing time series methods, comparing the Holt-Winters models and Box-Jenkins (ARIMA). To Moreira (1996) methods that apply time series seek to identify trends and seasonality on observations of demand over time, these checks raised in the series have a constant spacing in days, weeks, even years. This study showed that there was a reduction of 50% compared to the expected errors of the Box-Jenkins method to the Holt-Winters method.