Ajuste de distribuição de probabilidades de variáveis de custo fixo e variável na produção de suínos no estado de Santa Catarina

Abstract

Agribusiness is an extremely important sector in the world economy and in the brazilian economy, as it generates employment and income. Among the agribusiness sector products stand out sugar, coffee, soybean, corn, beef, swine, chicken and others. The research hereby presented is based on a real database from pig farming. Swine producers are constantly seeking for higher profits, in other words, they always try to obtain an expected maximum return on investment. Therefore it is necessary determine the behavior of the costs (fixed and variable) involved in the production. Since the random variable cost is a continuous variable, this research aimed to test the probability distributions for continuous variables, as Weibull, Rayleigh, Lognormal and Normal - identifying which of them have the best fit through the QQ-plot graph and p-values of the Kolmogorov-Smirnov test with probability 0,05. The Weibull distribution showed a better adjustment to variable total cost due to present p-value of 0,1867 in Kolmogorv-Smirnov test. For fixed cost total, the Rayleigh distribution adjusted the data better showing a p-value of 0,2910. It is suggested as a complement to this study, perform the same analysis with continuous probability distributions using deflated data, this way should be find the best distribution for the data.