本文以73至87學年度台灣地區國小教師人數為研究資料，比較指數平滑法、灰預測模型及模糊時間數列等模式，探討適用於國小教師人數短期預測之模型。經實證分析結果得知：（1）以灰預測模型與模糊時日數列預測台灣地區國小教師人數短期預測是可行的；（2）就兩種短期預測模戈來比較，GM（1,1）模型之預測誤差率為1.3%、模糊時間數列之預測誤差率為0.3%，模糊時間數列於樣本外有稍低之預測誤差率，是台灣地區國小教師人數短期預測之最佳模式；（3）依據89、90學年度之預測值來看，未來國小教師需求將會持續的增加。 The sample data consist of the number of elementary teachers in the Taiwan area from 1984 to 1998. In order to obtain more reliable short run forecasting results, this study has been compared to many different forecasting models.：exponential smoothing, grey forecasting, and fuzzy time series. The empirical result of analyzing here can be concluded as follows： The error of the GM（1,1）model and fuzzy time series are 1.3% and 0.3%, respectively. Therefore, both of the models are suitable for the short rum prediction of the issues；it does provide us evidence that the fuzzy time series is the best model among them. According to the forecasting values from 2000 to 2001it will continuously increase the demand for elementary teachers in the future.
教育與心理研究, 25(下),485-506 Journal of Education & Psychology