Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2001

This study modeled a time series data on corn for the period January 1989 to October 1997. Modeling was done for the purpose of generating a good forecast on the corn data. The four models used were autoregressive (AR) model, moving average (MA) model, autoregressive moving average (ARMA) model and autoregressive integrated moving average (ARIMA) model. The study underwent four main stages identification stage, estimation stage, diagnostic checking and forecast evaluation stage. EViews, which is the software used in this study, provided the needed tools for the four stages. Each stage underwent proper interpretation and proper techniques on how to handle the interpretations on hand. The study showed that the corn data is nonstationary, which means that its mean, variance and autocorrelation vary over time. A Log transformation was done to stabilize the nonconstant variance. It also underwent differencing to make the log transformed series stationary. The resulting model is given by (1-0.204B12) (1-B) LCorn = 1 + 0.211652) a , where LCorn ? Log(Corn). The forecast evaluation showed that the model was a predictive power (i.e.) mean absolute percent error or MAPE value of 3.93% and is good enough to use for forecasting.