| Summary: | 博士 === 國立臺灣大學 === 農業經濟學研究所 === 90 === The purpose of this study is mainly to compare the several forecasting models of the wholesale price for mangos in Taiwan in order to increase the precision of the forecast. The combining forecasts use econometric model, time series, and adaptive weighting respectively in the content, and another combining forecasts consist of three traditional regression models. Other than the stated combining forecasts in the records, this study also suggests to add adaptive weighting from the seasonal inaccuracy adjustment and three other seasonal regression models of combining forecasts to proceed with the wholesale price forecasting of mangos in Taiwan.
Covered from 1993 to 2001, this article utilizes the information of the wholesale prices and the trading volumes of mangoes and other important competitive fruits of Taipei Fruits and Vegetables Transportation and Sale Corporation to progress the mango wholesale price forecasting demonstration analysis. Taking statistics indexes such as the root mean square error (RMS), the root mean square percentage error (RMSPE), the mean absolute percentage error (MAPE), and the Theil’s inequality coefficient, etc. of the ten forecasting demonstration results in order to determine the evaluation of the forecasting value. The end result shows that the addressed added regression model from the seasonal adjustment in this study is capable of mounting the forecast efficiency among all statistics evaluation indexes compared to a unitary model and other traditional combining methods. Furthermore, it can step forward to increase the forecast precision, and appear to be quite stable in the simulation within the sampling period and in the forecasting performance outside of the sampling period.
|
|---|
