The application of weighted fuzzy time series to Interval forecasting

預測技術在決策過程中是不可或缺的重要工具。精確的預測可以提供決策者更多的資訊去做出正確的決策。傳統的點預測方法是目前使用最多的預測方式，其預測模式常需要較嚴格的基本假設，這使得預測模式的建構較為困難。而加權模糊時間數列模式並不需要強烈的基本假設，模式架構較傳統更為簡易，也提供決策者更多的選擇。本研究將傳統的加權模糊時間數列推廣為區間加權模糊時間數列。與常用的幾種區間模糊時間數列做比較，以預測每日台幣對美元的匯率的方式來探討幾種預測方法的效率評估與準確性。Forecasting technology has played an important role for the decision makers. Accurate forecasts can provide decision makers more information to make the right decisions. Currently, the most use of forecasts is the traditional point forecasting, whose forecasting model often requires strict assumptions, and this makes it more difficult to construct the forecasting model. Weighted fuzzy time series model does not require so strong assumptions, so the model construction is simpler than traditional ones. It also provides the decision makers more options. In this research, we promote the weighted fuzzy time series model to the interval weighted fuzzy time series model. And we compare it with some commonly used interval fuzzy time series models, to discuss their efficiency evaluations and accuracy by forecasting daily exchange rate for US Dollars to NT Dollars