本研究以台南市八處歷史文化財過去八年（一九九九年至二○○六年）之遊客人數作爲基礎資料，嘗試建構得以適切解釋遊客人數變化的時間序列模式。並依據確認後之模式進行遊客人數預測，藉以驗證模式建構之正確性。經由研究結果發現，在所建構的四項ARIMA模式之中，以ARIMA(3, 2, 1)(1, 2, 1)^12模式，較能適切反映及預測台南市歷史文化財遊客人數的增減趨勢。至於模式預測方面，雖然受到舉辦特定大型節慶活動之影響，形成預測配適值與實際值兩者之間有所誤差。惟就資料的整體變動方向而言，應仍可提供部份之觀察與思考。此亦同時顯示，在歷史文化財遊客人數模式建構的課題方面，依然具有諸多深入探討之空間與價值存在。 This study uses visitors' number of eight historic culture assets in Tainan city to analyze and explain the change of visitors' number during 1999 to 2006. To do this, we use time series analysis model to forecast visitors' number in the future. The result shows the ARIMA (3, 2, 1) (1, 2, 1)^12 model could explain and forecast the trend about the change of visitors' number of historic culture asset in Tainan city more powerful than another three ARIMA models. For model forecast, because of the impact of special festival activities, there are error exist between forecast value and real value. But for the change of whole data set, it still provides a part of observation and thinking. Furthermore, it also reveals the topic of model construction and practices for visitors' number still need to do more research in the future.
