| Summary: | 碩士 === 國立屏東科技大學 === 農企業管理系 === 89 === With getting the balance nutrition, pineapple comes to be the major fruit with rich nutrition and plays an important role in the food consumption structure. However a few of tangible and intangible factors make the price of pineapple volatile and unstable. During 1992-2000, the marketing price of pineapple changes from the highest of NT$72.36 to the lowest by NT$35.34, much more than two times the size. The volatile price will directly affect the producers and the consumers. The volatility of price drives this study to build a set of pineapple price forecast system.
This study use the grey prediction、back propagation neural networks, and combination of forecast model to forecast and analyze the marketing price of pineapple in this industry in Taiwan. By comparing the forecasting accuracy from these three models, the best efficient one was selected to forecast the agricultural price faster and more exact.
Comparing to the forecasting accuracy or the absolute forecasting error, the empirical results showed that the combination of forecast model is more efficient than the grey prediction and back propagation neural networks. But the combination of forecasting A has the highest forecasting accuracy because the model restriction is less than the others are. In average, their forecasting ability has reached an efficient level. No matter the single model of the grey prediction or the back propagation neural networks or the combination of forecast model enables to forecast the marketing price of pineapple effective
|
|---|
