區間時間數列預測及其效率評估

點預測為目前使用最多之預測陳述，其效率評估亦多以最小平方和誤差(minimum of sum of square errors)為主。每日或月的經濟或財金指標預測是點預測最常見的例子。但是隨著區間時間數列真正需求與軟計算(soft computing)科技的發展，區間計算與預測愈來愈受重視。本文提出幾種區間時間數列預測的方法，並研究其效率評估。最後我們以影響經濟作物的天氣預測，作實證研究分析。考慮在無參數條件下，幾種預測方法作效率評估與準確性探討。天氣預測是區間預測的例子，建立合適的的區間預測方法與效率評估，對各研究領域將會有莫大的幫助。

關鍵字

區間時間數列；天氣預測；區間預測；效率評估

並列摘要

The point forecasting stated for present use most forecasts, its efficiency evaluation also many by least squares and error (minimum of sum of square errors) primarily. Either the month economy or the wealth gold target forecasting is every day the point forecasting the most common example. But along with interval time series real demand and soft computation (soft computing) technical development, the interval computation and the forecasting receive more and more take seriously. This article proposed that several interval time series forecasting's method, and studies its efficiency evaluation. Finally we affect the industrial crop the weather forecasting, makes the empirical study analysis. The consideration under the non-parameter condition, several forecasting techniques makes the efficiency evaluation and the accurate discussion. The weather forecasting is the f interval forecasting example, establishes the appropriate interval forecasting technique and the efficiency evaluation, will have the greatest help to each research area.