Application of ARIMA Models in Soybean Series of Prices in the North of Paraná

This study aims to analyze the series of daily prices of soybeans in the North of Parana, which includes the timeline of the years 2000 (from January) to 2011 (until October) and describe their behavior with short-term forecasts. That is, verify if the temporal dynamics of the variable is better explained by a process: Auto-regressive of order p, moving Average of order q, Auto-regressive and moving Average of order p,q; Auto-regressive, Integrated and moving Average and of order p, d, q; or seasonal ARIMA of order (p, d, q), (P, D, Q). The Box and Jenkins methodology (1976) was used to answer this question. The manipulation of data was based on graphical analysis and statistical tests of the methodology itself, through which, it was observed that the ARIMA (5,0,0) or simply AR (5), responded as the best model among the set of models tested to predict the prices of soybeans.