Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2007

This study was conducted to develop short-term forecasts utilizing the time series data on the monthly average domestic prices of copra in the Philippines for January 1991 to December 2003. Also, this study explored the capacity of Box and Jenkins analysis to model the variable being studied. The specific objectives of this study were to describe first the behavior of the data series, then develop a univariate model to be used in drawing a one-year forecast and lastly, to check the validity of the model by comparing the forecasted with the actual values. Box and Jenkins analysis requires that the data series is stationary, that is, it should be independent of time, have a non-constant mean and non-constant variance. Thus, the raw data series was thoroughly investigated of its stationary. Results showed that it is nonstationary, as a result, appropriate transformation were applied. Taking the second difference of the natural log-transformed series achieved stationary. It was then used in deriving a model. And the final model identified using the transformed series was a MA (2) model. The accuracy of the resulting model was tested by checking it residuals and by comparing the forecasted with the actual values. In addition, the result given by the computed value for mean absolute percentage error (MAPE) indicated that the model has a high predictive power.