多變量模糊時間數列模式之應用--以臺灣地區高職教師人數之預測為例

Application of Multivariable Fuzzy Time Series Models: Forecasting Numbers of Teacher in Vocational High School in Taiwan Area

本文提出以多變量模糊時間數列模式??二因子模式、引導式模式、馬可夫模式，建構臺灣地區高職教師人數之預測模式。實證結果得知在有限資料筆數下，適應性類神經模糊推論系統（簡稱ANFIS）預測結果較佳，多變量模糊時間數列模式次之，灰色預測模式第三。而在方法操作上，以多變量模糊時間數列引導式模式最為簡易。

In this paper， we apply multivariable fuzzy time series models to forecast numbers of teachers in vocational high school. The empirical result of analyzing here can be concluded as follows: The error of Adaptive Neuro-Fuzzy Inference Systems (ANFIS) is the smallest， multivariable fuzzy time series models is the second smallest， and grey forecasting is the third smallest. The multivariable fuzzy time series Heuristic model is the easiest method to follow.